# Bynn Intelligence: the full text of the main pages > Bynn Intelligence builds enterprise AI through frontier research, proprietary models, autonomous systems, and AI development and integration. Generated from the content the pages at https://bynn.com/ render. The index is at https://bynn.com/llms.txt. Each result keeps its evidence type, conditions and source. ## For coding agents To integrate Bynn, paste this line into your coding agent. It installs the Bynn skills and the Bynn MCP server, then works from the documentation and the OpenAPI spec: ```text Set up Bynn, ZOREAL, DecisionNode, Agemin and Detector24 for me: fetch https://bynn.com/agent-setup.md and follow it. ``` - Setup instructions (agent-setup.md), on our site: https://bynn.com/agent-setup.md - Setup instructions (agent-setup.md), source: https://github.com/Bynn-Intelligence/skills/blob/main/agent-setup.md - Skills repository: https://github.com/Bynn-Intelligence/skills - Bynn API documentation: https://docs.bynn.com - Documentation index for agents: https://docs.bynn.com/llms.txt - OpenAPI spec: https://api.bynn.com/openapi.json - MCP server: https://mcp.bynn.com (documentation: https://docs.bynn.com/docs/bynn-mcp) - API keys: https://dashboard.bynn.com - Bynn API (https://docs.bynn.com): Document fraud detection for PDF and image documents, identity verification sessions, AML screening, content moderation, face search and webhooks. Use it when a task mentions KYC, identity, onboarding or a document check. - ZOREAL (https://zoreal.com): Human trust for digital interactions: identity verification and document forensics in ZOREAL Verify, verified meetings, sign-in and signing. - DecisionNode (https://decisionnode.com): A decision model and inference service that returns decisions connected systems can act on. - Agemin (https://agemin.com): AI for age assurance, built on Bynn's facial age-estimation research. Use it for a self-serve age gate; it has its own web SDK (@bynn-intelligence/agemin-sdk), not the Bynn one. - Detector24 (https://detector24.ai): Specialist models that examine what AI can create and conceal in images, video, audio and text. ## Bynn Developer Documentation: Build With Your Coding Agent URL: https://bynn.com/docs Developer documentation for every Bynn platform: the Bynn API, ZOREAL, Agemin, Detector24 and DecisionNode. One prompt onboards your coding agent. ### Build with Bynn One prompt sets your coding agent up with every Bynn platform: the skills, the MCP servers and the API spec. ### Bynn API Document fraud detection for PDF and image documents, identity verification sessions, AML screening, content moderation, face search and webhooks. - Get started: https://docs.bynn.com/docs/getting-started - API reference: https://docs.bynn.com/reference - Webhooks: https://docs.bynn.com/docs/bynn-webhooks ### ZOREAL Human trust for digital interactions: identity verification and document forensics in ZOREAL Verify, verified meetings, sign-in and signing. - ZOREAL API documentation: https://zoreal.com/documentation - Verify: document fraud detection: https://docs.bynn.com/docs/document-fraud-detection - Login with ZOREAL SDK: https://github.com/Bynn-Intelligence/zoreal-oauth2-js ### Agemin AI for age assurance, built on Bynn's facial age-estimation research. - Documentation: https://agemin.com/docs - Web SDK: https://github.com/Bynn-Intelligence/agemin-sdk - API reference: https://docs.bynn.com/reference/postageverificationliveness ### Detector24 Specialist models that examine what AI can create and conceal in images, video, audio and text. - Documentation: https://detector24.ai/docs - Model catalogue: https://detector24.ai/model-catalogue - AI-generated image model: https://detector24.ai/model/ai-generated-image ### DecisionNode A decision model and inference service that returns decisions connected systems can act on. - Quickstart: https://decisionnode.com/docs/quickstart - API reference: https://decisionnode.com/docs/api - MCP server: https://decisionnode.com/docs/mcp ### Zora A mission executive and autonomy engine for autonomous systems, demonstrated in simulation. Zora is shown in simulation and has no public developer documentation. ### Documentation by task - Verify identities (Bynn API): https://docs.bynn.com/docs/bynn-onboarding-process - Detect document fraud (ZOREAL Verify, through the Bynn API): https://docs.bynn.com/docs/document-fraud-detection - Add a decision engine (DecisionNode): https://decisionnode.com/docs/quickstart - Gate by age (Agemin): https://agemin.com/docs - Detect AI-generated media (Detector24): https://detector24.ai/docs - Screen for AML (Bynn API): https://docs.bynn.com/docs/aml-check - Sign in with ZOREAL (ZOREAL, for React): https://github.com/Bynn-Intelligence/zoreal-oauth2-react - Moderate content (Bynn API): https://docs.bynn.com/reference/postmoderationinfer - Verify without writing code (Bynn API): https://docs.bynn.com/docs/no-code - Receive results by webhook (Bynn API): https://docs.bynn.com/docs/bynn-webhooks ## Enterprise AI & Frontier Research | Bynn Intelligence URL: https://bynn.com/ Bynn Intelligence builds enterprise AI through frontier research, proprietary models, autonomous systems, and AI development and integration. ### Frontier AI for the enterprise. Bynn Intelligence is a frontier AI research lab developing enterprise AI, from proprietary models and decision engines to autonomous systems. ### Research with independently measured results. - #1 for true-positive rate in NIST's Child Online Safety evaluation for ages 13-16. True-positive rate, ages 13-16: 72.53%. Bynn's facial age-estimation model, which powers Agemin, achieved a 72.53% true-positive rate: the proportion of people aged 13 to 16 correctly estimated within that age band. Measured by NIST in its ongoing evaluation program. NIST evaluation is not certification or endorsement. The result measures age estimation within the 13 to 16 band. It does not measure a reduction in online harm. Evidence type: Independent evaluation. Source: NIST FATE Age Estimation & Verification. Submission bynn-000. https://pages.nist.gov/frvt/html/frvt_age_estimation.html (NIST results table last updated 29 September 2026). ### Intelligence becomes useful when it can act. Our research turns into systems that carry out missions, make decisions and examine evidence. - Zora: Autonomous search and rescue: Zora's mission executive coordinates an aerial search, while its autonomy engine responds to what it observes. (Zora Air: autonomous mission execution in simulation.) https://bynn.com/products/zora - DecisionNode: Deciding what happens next: A package reaches a sorting line. DecisionNode evaluates its condition and label, selects a route, and returns a decision the line can execute. (Published workflow example, not a customer deployment.) https://bynn.com/products/decisionnode - Detector24: Claim photos checked for AI edits: Detector24's image-forgery model examines a photo submitted with an insurance claim and marks the region that was edited. (Example analysis of an AI-edited image.) https://bynn.com/products/detector24 - ZOREAL: Is that really your CFO on the call? ZOREAL Meet lets participants prove they are verified, live people, and checks camera video for face-swap deepfakes. (Example meeting with fictional participants.) https://bynn.com/products/zoreal - Agemin: Age checks from a face or an email: Agemin estimates age from a face or an email address, so a service can open the experience that suits the person. https://bynn.com/products/agemin ### Research, put to work. Five AI products across digital trust, forensics, age assurance, decision intelligence and autonomy. - ZOREAL (Digital trust): Human trust for digital interactions: identity verification and document forensics in ZOREAL Verify, verified meetings, sign-in and signing. https://bynn.com/products/zoreal - Agemin (Age assurance): AI for age assurance, built on Bynn's facial age-estimation research. https://bynn.com/products/agemin - Detector24 (Media forensics): Specialist models that examine what AI can create and conceal in images, video, audio and text. https://bynn.com/products/detector24 - DecisionNode (Decision intelligence): A decision model and inference service that returns decisions connected systems can act on. https://bynn.com/products/decisionnode - Zora (Autonomy): A mission executive and autonomy engine for autonomous systems, demonstrated in simulation. https://bynn.com/products/zora ### Enterprise AI. From research to deployment. Bring us the challenge. We develop the models, decision engines, autonomous workflows and infrastructure to address it. From AI consulting and custom development to implementation and integration, we work with enterprises to turn a business need into a working system. - AI consulting and strategy: Identify where AI can make a difference, define the problem and establish what the system needs to achieve. https://bynn.com/enterprise-ai#consulting - Custom AI development: Develop models, decision engines and autonomous systems around your data, workflows and requirements. https://bynn.com/enterprise-ai#custom-development - AI implementation and integration: Connect AI to your existing systems so it can use the right information and act within the workflows that matter. https://bynn.com/enterprise-ai#implementation - Autonomous workflows: Automate the steps around a decision, from collecting documents and assessing evidence to triggering the next action, with optional human review when needed. https://bynn.com/enterprise-ai#workflows - Inference and infrastructure: Bring together model serving, inference and GPU resources as part of the stack an AI system needs to operate. https://bynn.com/enterprise-ai#infrastructure ### Understand the evidence. Automate the work around it. We build tools to identify signs of altered, fabricated and AI-generated evidence, from financial documents to images submitted with a claim. - Document fraud detection: Examine submitted documents for signs of tampering, fabrication and hidden changes. https://bynn.com/document-fraud-detection#document-fraud - AI document collection: Automate document requests and follow-up across languages, helping the right information reach the right workflow. https://bynn.com/document-fraud-detection#document-collection - Agentic fraud reasoning: Apply AI reasoning to fraud assessment within a wider evidence-review workflow. https://bynn.com/document-fraud-detection#fraud-reasoning - Image and deepfake detection: Assess images and other media for signs of generation or manipulation using specialist models. https://bynn.com/document-fraud-detection#media-detection ### Enterprise AI for your industry. Document forensics, identity verification, AI workflows and decision engines, mapped to the work of each sector. - Banking: Bring document forensics, identity verification and AI workflows into onboarding, lending and financial-document review. Examine the bank statements, payslips and company records behind an application, and automate the work of collecting them. https://bynn.com/industries/banking - Insurance: Automate document intake and examine claim evidence with AI. Multilingual collection, document fraud detection, image analysis and agentic fraud reasoning work together to support better claims decisions. https://bynn.com/industries/insurance - Energy and utilities: Explore AI for infrastructure inspection and the operational work around it, including autonomy research for power-line inspection. Automate the documents and decisions behind new connections, field work and customer accounts. https://bynn.com/industries/utilities - Retail: Apply image analysis, content moderation and age assurance to digital commerce and customer interactions. Check seller documents, listing images and the photos behind a refund claim. https://bynn.com/industries/retail - Healthcare: Explore identity verification, document collection and administrative automation that help teams spend less time coordinating information. The work stays administrative: patient identity, referrals, claims paperwork and staff credentials. https://bynn.com/industries/healthcare - Telecommunications: Bring identity verification, document checks and AI workflow automation to customer onboarding and service operations. Examine the scam messages and cloned voices that reach customers through your network. https://bynn.com/industries/telecommunications - Public services: Explore AI for document collection, verification and administrative workflows across languages. Reduce the manual coordination between an application and a decision, with people in charge of the outcome. https://bynn.com/industries/public-services - Smart cities: Explore decision engines for urban infrastructure, from traffic-signal decisions to automated operational workflows. DecisionNode provides decisions that connected systems can act on. https://bynn.com/industries/smart-cities - Public safety: AI for the challenges facing emergency responders, police and border-protection teams. Explore document and media authenticity analysis, evidence-review workflows and autonomy research for search and rescue. https://bynn.com/industries/public-safety - Defense: Explore DecisionNode's application in drone-detection and situational-awareness systems, including systems designed to recognize groups of drones. Decisions for detection and awareness, with no weapons capability. https://bynn.com/industries/defense ### From understanding threats to building intelligence. Bynn's foundations are in fraud detection and forensic intelligence: understanding how deception works and how to recognize it. Building our own AI models and systems has expanded both our expertise and our ambition. ### What could your organization do with AI? Bring us a challenge, a partnership opportunity, or an interest in the company we are building. For enterprises, the conversation starts with your needs. For partners, it starts with what we can create together. For investors, it starts with the research, products and ambition behind Bynn Intelligence. Or write to us at hello@bynn.com ### Bynn Intelligence: Are you with us? A forged document. A fire crossing a hillside. A search where every second matters. The world needs intelligence that can meet the moment. Bynn Intelligence is a frontier AI research lab building proprietary models, decision engines and autonomous systems. Transcript: A forged document. A fire crossing a hillside. A search where every second matters. The world needs intelligence that can meet the moment. We are Bynn Intelligence. A frontier AI research lab building proprietary models, decision engines, and autonomous systems. ZOREAL is building a trust layer for the internet. Verifying identity. Bringing proof of humanity to digital interactions in a world where seeing is no longer believing. Our forensic AI examines documents and media for signs of manipulation and fraud. Agemin brings our age-estimation research into online safety. Its model ranked first for true-positive rate in NIST's Child Online Safety evaluation. DecisionNode turns information into decisions that connected systems can act on. From enterprise workflows to robotics, it determines what happens next. Inside Zora, it provides tactical decisions. Zora is our mission executive and autonomy engine, already demonstrating autonomous flight in simulation. Our ambition reaches further. Search and rescue. Wildfire response. Energy infrastructure. Defense reconnaissance. Helping identify mines and unexploded ordnance after conflict. Intelligent systems that take on difficult work and help protect people. We have built working products and the research behind them. Now, we are raising capital to accelerate commercial growth. Reach more customers. Expand deployments. Bring more of our research into the world. The next chapter is scale. Bynn Intelligence. Frontier AI for the enterprise. Are you with us? ## Enterprise AI Development & Implementation URL: https://bynn.com/enterprise-ai Enterprise AI from a frontier AI research lab: AI strategy consulting, custom AI development, AI implementation and integration, AI workflows and inference. ### Enterprise AI from the lab that builds its own models. Bring us the challenge. We develop the models, decision engines, autonomous workflows and infrastructure to address it. ### One team from the first question to the running system. From AI consulting and custom development to implementation and integration, we work with enterprises to turn a business need into a working system. An engagement can draw on our portfolio, our proprietary models, or technology developed around a specific challenge, with inference and GPU infrastructure to support it. Because we build our own models, we can tell you when an existing one will do and when it will not. Not every engagement needs a new model. Every engagement needs the right one. ### AI strategy that ends in a decision, not a slide deck. Question: Where should we start? Most organizations have more possible uses for AI than they can pursue. AI strategy consulting with Bynn starts from the work: which decisions are slow, which processes depend on people moving information between systems, and where better evidence would change an outcome. We help you choose where to begin, define the problem precisely and agree what the system must achieve, including how it will be measured and where people stay in the loop. - Find the opportunity: Map processes and decisions where AI can make a measurable difference. - Define the problem: Turn a broad ambition into a specific task with inputs, outputs and limits. - Set the bar: Agree how success is evaluated before anything is built. ### Custom AI, built around the work you do. Question: Can you build AI around our data and our problem? Custom AI development at Bynn draws on the same research behind our portfolio: decision models, forensic models, age estimation and autonomy. When an existing model fits, we adapt it. When a problem is new, we develop the model, decision engine or autonomous system it needs. Your data and your requirements shape the system: what it must recognize, what it must decide, how fast, and how it explains itself. - Models: Specialist models trained or adapted for your inputs, from documents and images to text and sensor data. - Decision engines: Systems that return typed decisions with a confidence, ready for software to act on. - Autonomous systems: Agents and mission executives that coordinate several steps toward a goal. ### Bring AI into your existing systems. Question: How do we turn a pilot into something our organization can use? A model only creates value when it is connected to the systems where work happens. Our AI implementation and integration services connect AI to your data sources, applications and workflows, so it receives the right information and its output reaches the place it is needed. Integration includes evaluation: testing the system on your own data, agreeing the thresholds at which it acts alone, and deciding what happens when it is unsure. - Connect: Integrate with the systems your teams already use, through APIs and existing workflows. - Evaluate: Test against your own cases before the system acts on live work. - Hand over: Agree who operates the system and how it is monitored once it runs. ### Autonomous workflows for complex operations. Question: Can AI do the work across several steps and systems? An autonomous workflow does more than answer a question. It interprets what arrives, decides what to do, coordinates the next step and acts, across several systems if needed. In insurance, that means requesting the documents a claim needs in the claimant's language, following up on what is missing, examining the evidence and routing the claim. Workflows run autonomously by default, with optional human review whenever an agent is unsure how to proceed. - Interpret: Read documents, images and messages as they arrive. - Decide: Choose the next step with DecisionNode or a model built for the task. - Act: Send the request, update the record, route the case, or ask a person. ### Inference and infrastructure, as part of the system. Question: How will it run, and who provides the stack? Every AI system needs somewhere to run. Bynn brings together model serving, inference and GPU resources as part of the stack an engagement needs, alongside the models and workflows themselves. Deployment environments and operating terms are agreed for each engagement, around your requirements for data location, security and scale. - Model serving: Serve the models and decision engines a system depends on. - Inference: Run inference for workflows and applications, including DecisionNode's inference service. - GPU resources: Provide the compute an engagement needs to train and run its models. ### How an engagement works Every engagement is scoped to the challenge, so the shape varies. These are the stages most of them pass through. - Discover: We learn how the work is done today: the people, systems, documents and decisions involved. - Define: We agree the problem, what the system must achieve, how it will be evaluated and where people review its work. - Build: We adapt existing models or develop new ones, and assemble the decision logic and workflow around them. - Integrate: We connect the system to your data and applications and evaluate it on your own cases before it acts on live work. - Operate: The system runs on agreed infrastructure, with monitoring and the human review points you defined. Timelines, service levels and commercial terms are agreed for each engagement. We do not publish generic delivery promises. ### Built on technology we already run. An engagement can start from Bynn's portfolio: ZOREAL for identity, trust and document forensics, Agemin for age assurance, Detector24 for media forensics, DecisionNode for decisions and Zora for autonomy research. That means less to build from nothing, and models already at work in our own products. ### Questions enterprise buyers ask Q: Do we need to train a new model? A: Often not. Many problems can be solved by adapting an existing model or combining our portfolio's models in a workflow. We develop a new model when the problem calls for one. Q: Can you work with our existing systems? A: Yes. AI implementation and integration is part of every engagement: connecting to the data sources, applications and workflows your teams already use. Q: How do you decide when AI acts on its own? A: Together with you, during definition and evaluation. We agree the confidence thresholds at which a system acts alone and the cases that always go to a person. Q: Do you provide the infrastructure as well? A: We can. Model serving, inference and GPU resources can be part of an engagement. Where and how the system runs is agreed around your requirements. Q: Is this consulting or software? A: Both, depending on what the challenge needs. AI strategy consulting is a common starting point, and it leads into development, integration and operation by the same team. Q: Which industries do you work with? A: Our deepest work is in banking and insurance, especially document fraud and claims. Our capabilities also apply to energy and utilities, retail, healthcare, telecommunications, public services, smart cities, public safety and defense. Q: How do we start? A: Book a meeting and bring a challenge. The first conversation is about your work, not our products. ### Bring us the challenge. Tell us about the decision, the process or the problem. We will tell you honestly what it would take to address it with AI. ## Enterprise AI Products and Portfolio URL: https://bynn.com/products ZOREAL, Agemin, Detector24, DecisionNode and Zora: the portfolio behind Bynn's enterprise AI, from digital trust and forensics to decisions and autonomy. ### Five products, one research lab. ZOREAL for digital trust, Agemin for age assurance, Detector24 for media forensics, DecisionNode for decisions and Zora for autonomy. ### The Bynn portfolio Bynn Intelligence is the AI research lab behind ZOREAL, Agemin, Detector24, DecisionNode and Zora. Across digital trust, forensics, age assurance, decision intelligence and autonomy, our portfolio puts research to work. Each has its own focus, team and identity. Together, they connect our research with commercial products, specialist platforms and new applications of AI. Their maturity differs, and each page says where it stands. - ZOREAL (Digital trust, Available): Human trust for digital interactions: identity verification and document forensics in ZOREAL Verify, verified meetings, sign-in and signing. https://bynn.com/products/zoreal; own site https://zoreal.com - Agemin (Age assurance, Available): AI for age assurance, built on Bynn's facial age-estimation research. https://bynn.com/products/agemin; own site https://agemin.com - Detector24 (Media forensics, Available): Specialist models that examine what AI can create and conceal in images, video, audio and text. https://bynn.com/products/detector24; own site https://detector24.ai - DecisionNode (Decision intelligence, Available): A decision model and inference service that returns decisions connected systems can act on. https://bynn.com/products/decisionnode; own site https://decisionnode.com - Zora (Autonomy, Research): A mission executive and autonomy engine for autonomous systems, demonstrated in simulation. https://bynn.com/products/zora ### How the pieces connect The portfolio is not a list of separate tools. Several entries rely on each other. - Zora uses DecisionNode: DecisionNode is Zora's System One: it answers the tactical questions about each camera frame, and Zora coordinates the mission and its execution. - DecisionNode draws on Agemin and Trinity: Boosters add specialist evidence to a decision. Two come from the portfolio: Human Age from Agemin, and AI-generated image detection from Detector24's Trinity. - Forensics in ZOREAL Verify and Detector24: ZOREAL Verify, which includes PDFChecker, examines documents and images; Detector24 brings specialist models for images, video, audio and text. Each keeps its own focus, and both are available. - Identity moves to ZOREAL Verify: Bynn's full identity-verification product, with document and biometric checks and onboarding, is moving to ZOREAL Verify as a complete product. ### And the enterprise practice that connects them. We also work directly with enterprises, taking AI from a business challenge through research, development, integration and the infrastructure needed to run it. An engagement can draw on any part of the portfolio, or on technology developed for the challenge itself. ### Where Bynn's earlier offerings live now Bynn.com used to list its services one by one. None of them has been dropped. Here is where each one lives now. #### Identity and age - Identity verification: Moving to ZOREAL Verify as a complete product. - NFC verification: Available from Bynn, alongside identity verification. - Face authentication: Available from Bynn, alongside identity verification. - Proof of address verification: Available from Bynn, with identity verification and document collection. - Age verification: Agemin. #### Compliance and risk - AML screening: Available from Bynn. - Know your business: Available from Bynn. - Bank verification: Available from Bynn. - Customer risk assessment: Available from Bynn, alongside agentic fraud reasoning. - Phone verification and validation: Available from Bynn. - Proxy and anonymous-IP detection: Available from Bynn. #### Documents and evidence - Document fraud detection: Bynn's capability page; moving into ZOREAL Verify. - Document fraud detection tool: PDFChecker, part of ZOREAL Verify. - AI document collection: Bynn's autonomous, multilingual collection workflows. - Agentic fraud reasoning: Bynn, within document fraud and claims workflows. - Autonomous claims workflows: Bynn, on the insurance industry page. - Utility-bill verification: Document forensics, available from Bynn. - Accounts-payable fraud: Document fraud detection applied to supplier invoices. #### Media and content - AI-generated image, video and audio detection: Detector24. - Inappropriate content and media moderation: Detector24. - Scam and fraud SMS and text moderation: Detector24 text models. #### Industry solutions - Insurance fraud and insurance fraud detection: Insurance industry page. - Loan underwriting and fintech: Banking industry page. - Merchant onboarding and e-commerce: Banking and retail industry pages. - Know your patient: Healthcare industry page. - Tenant screening and real estate: Document fraud detection and identity verification, applied to rental and property checks. - Crypto, gambling, dating, human resources, transport, web hosting, proxy networks and music: Applications of identity verification with ZOREAL, and age checks with Agemin where age applies. ## ZOREAL: Identity Verification and Digital Trust URL: https://bynn.com/products/zoreal ZOREAL brings Bynn's identity verification and document forensics together in ZOREAL Verify, alongside verified meetings, sign-in and signing. ### Human trust for digital interactions. Takes in: A person, their government ID, a document or a live call. Returns: A verifiable answer about who is there, and what a document hides. Status: Available. ZOREAL Verify, ZOREAL Meet and ZOREAL Sign are available. Verify's checks run through Bynn today, and its integrated dashboard is coming soon. ZOREAL ID enrolment runs on Android, with the iOS app in testing. ZOREAL Mark is coming soon. Own site: https://zoreal.com Pricing: From $4.95 per user per month (Premium), as published at https://zoreal.com/pricing ### Trust between people, established before it is assumed. ZOREAL establishes who is on the other side of a digital interaction: that they are a real person, that they are present, and who they are. ZOREAL brings together identity verification, document forensics, human presence, verified sign-in and signing. From verifying a customer's identity to assessing the documents they submit, it helps organizations establish trust across digital interactions. Bynn's full identity-verification product is moving to ZOREAL Verify: document and biometric identity checks and the onboarding workflow around them. It joins ZOREAL as a complete product, not as a sign-in feature. The questions ZOREAL answers have become harder as generated faces and cloned voices improve. Is this really the colleague asking for a payment? Is the applicant the person on the document? ZOREAL is built to give a verifiable answer, and to say exactly what that answer does and does not prove. ### A verifiable answer to who is on the call. In a ZOREAL Meet call, each participant can prove they are a real, enrolled, live person, and share their verified legal name only when they choose to. ### What ZOREAL does - ZOREAL Verify (Available): Identity verification and document forensics, including PDFChecker: document and biometric identity checks inside an onboarding flow, and analysis of PDFs, document images and images for signs of alteration, fabrication and AI generation. The checks run through Bynn today; the integrated dashboard is coming soon. - ZOREAL Meet (Available): Video and voice meetings where a participant can prove they are a real, enrolled, live person, either present or present with their verified legal name. In standard rooms, Meet also checks camera video for face-swap deepfakes. - ZOREAL Book (Available): Scheduling pages inside Meet. A guest picks a free slot that never clashes with the host's own calendar, and a booking with a room becomes a scheduled Meet call. - ZOREAL Identify (Available): A sign-in button on standard OpenID Connect. The website receives a stable pseudonymous identifier for a verified human and how strongly and how recently that was proven. Identity details are shared only when the person consents. - ZOREAL ID (Available): An identity wallet that turns a physical government ID into a credential on the owner's phone. It reads the document's security chip over NFC and binds it to a liveness capture. - ZOREAL Sign (Available): Signature dossiers prepared, sent and signed in the browser, with every step recorded in an audit log that is sealed, attested and timestamped when the dossier completes. - ZOREAL Mark (Coming soon): A way for a verified person to vouch for what they post, so anyone can check that a real human verified by ZOREAL stood behind exactly this text, on this page, at this time. It proves who vouched, not who wrote it. ### ZOREAL Verify: Is the document genuine, and is the person who they say they are? ZOREAL Verify brings Bynn's document and image forensics, including PDFChecker, together with Bynn's full identity-verification product. It examines how a document or image may have been created or changed, and confirms that the person in front of the camera is live and matches their government ID. Status: Available. Available. The checks run through Bynn today, and the integrated ZOREAL Verify dashboard is coming soon. Bynn's identity verification and document fraud detection are moving into Verify and stay available from Bynn in the meantime. #### Document and image forensics Look deeper into a document. PDFChecker, part of ZOREAL Verify, examines PDFs and document images for signs of alteration and fabrication, and brings its findings together in an authenticity report. - Structure and metadata: How the file was produced and saved: its internal structure, producing software, timestamps and revision history, and the traces an edit leaves behind. - Edits and tampering indicators: Changed amounts, replaced payment details and altered text, including revisions that are invisible on the rendered page, and fonts, spacing or graphics that do not match across fields. - Signatures: Inspection of embedded digital signatures, and whether a signed document has remained intact since it was signed. - Classification: Identifies the type of document so the right checks apply to it. - Images: AI-generated images and signs of manipulation within an otherwise genuine image. - Authenticity report: Findings gathered into one report that shows what deserves attention, where it appears and why it matters, for a reviewer to act on. #### Identity verification Bynn's full identity-verification product is moving to ZOREAL Verify as a complete product: document and biometric identity checks and the onboarding workflow around them. - Identity document checks: A government ID is read, including its machine-readable zone, and checked for authenticity and signs of tampering. - Face match: The person's face is matched against the photo on their document. - Liveness: A liveness check confirms that a live person is present for the capture. - Onboarding: The checks run inside a configurable onboarding flow, so each customer journey asks for what it needs. Identity verification is not ZOREAL Identify. Identify is a sign-in button for a verified human and is not KYC; Verify's identity checks are the onboarding product. ### How ZOREAL connects - Built on Bynn's research: ZOREAL's liveness capture and deepfake detection are developed in-house at Bynn, alongside its forensic and age-estimation models. ZOREAL turns that work into the trust layer of a digital interaction. - ZOREAL Verify and Detector24: ZOREAL Verify examines documents, images and identities. Detector24 keeps its own focus: specialist models for images, video, audio and text. - Agemin: Agemin keeps its own age-assurance offering, including ID verification for age checks. The identity product moving to ZOREAL Verify does not merge the two. ### Bynn offerings moving to ZOREAL These services were offered on bynn.com. Their move into ZOREAL Verify is planned, and they stay available from Bynn in the meantime. - Identity verification: Planned move to ZOREAL Verify as a complete product. - Document fraud detection: Moving into ZOREAL Verify. - Document fraud detection tool: PDFChecker, part of ZOREAL Verify. ## Agemin: Facial Age Estimation and Age Assurance URL: https://bynn.com/products/agemin Agemin brings Bynn's facial age-estimation research, #1 for true-positive rate in NIST's Child Online Safety evaluation for ages 13-16, into age assurance. ### AI for age assurance. Takes in: A face on camera, or an email address. Returns: An age decision the service can act on, with a step-up when needed. Status: Available. Available. NIST's evaluation is ongoing, and its tables change as new submissions are added. Own site: https://agemin.com Pricing: From $99 a month, monthly minimum commitment (Core), as published at https://agemin.com/pricing ### Age assurance that fits the moment it is needed. Agemin helps digital services assess age and build age-appropriate experiences, using a facial age-estimation model evaluated in NIST's ongoing testing program. Its offering spans facial and email-based age estimation, identity verification and age-related content tools. Agemin brings Bynn's facial age-estimation research, evaluated through NIST's ongoing testing program, into practical age assurance. The same model ranked first for true-positive rate in NIST's Child Online Safety evaluation for ages 13-16. Different services need different levels of certainty. A quick estimate can open the right experience; a document check can follow when the stakes require it. ### The right experience, from a quick estimate. Different services need different levels of certainty. Agemin starts with an estimate and asks for more only when the stakes require it. ### What Agemin does - Facial age estimation: Estimates a person's age from their face on camera, so a service can decide which experience to open without asking for a document first. - Email age estimation: Assesses age from an email address, for flows where a camera step is not wanted. - ID verification: A document-based check for the moments when an estimate is not enough. - Creator moderation: Age-related checks for the people who publish on a platform, not only those who watch. - Content age monitoring: Age-related monitoring of content a platform hosts, to support age-appropriate experiences. ### Independent evidence - #1 for true-positive rate in NIST's Child Online Safety evaluation for ages 13-16. True-positive rate, ages 13-16: 72.53%. Bynn's facial age-estimation model, which powers Agemin, achieved a 72.53% true-positive rate: the proportion of people aged 13 to 16 correctly estimated within that age band. Measured by NIST in its ongoing evaluation program. NIST evaluation is not certification or endorsement. The result measures age estimation within the 13 to 16 band. It does not measure a reduction in online harm. Evidence type: Independent evaluation. Source: NIST FATE Age Estimation & Verification. Submission bynn-000. https://pages.nist.gov/frvt/html/frvt_age_estimation.html (NIST results table last updated 29 September 2026). ### How Agemin connects - Bynn's age-estimation research: Agemin is where Bynn's facial age-estimation model meets real services. Research results, such as the NIST evaluation, apply to the model that powers it. - A DecisionNode booster: Human Age, the Agemin booster, lets DecisionNode add an age estimate to a decision when a workflow needs one. - ZOREAL: Bynn's enterprise identity verification is moving to ZOREAL Verify. Agemin keeps its own ID verification for age checks; the move does not merge the two. ### Bynn offerings that now live with Agemin Age verification was offered on bynn.com. It now lives with Agemin. - Age verification: Agemin. - Age verification software: Agemin. ## Detector24: AI-Generated Image and Deepfake Detection URL: https://bynn.com/products/detector24 Detector24 brings specialist detection models to images, video, audio and text: AI-generated images, image forgery, deepfakes and moderation. ### Examine what AI can create and conceal. Takes in: Images, video, audio and text. Returns: A separate signal from each specialist model, regions marked. Status: Available. Available. Catalogue entries are published by Detector24; each model's scope and limits are documented on its own page. Own site: https://detector24.ai Pricing: From $99 a month, monthly minimum commitment (Core), as published at https://detector24.ai/pricing ### One question per model, answered well. Detector24 brings specialist detection models to images, video, audio and text, with distinct signals for different kinds of evidence. Detect signs of AI-generated images, image forgery and deepfakes, alongside a broader range of moderation and analysis capabilities. Trinity focuses on identifying fully AI-generated images. Other specialist models address edited imagery and manipulated media, giving applications distinct signals for different kinds of evidence. A photo can be real and edited, or generated from nothing. Treating those as separate questions gives a reviewer a clearer answer than one blended score. ### Separate questions, separate signals. A photo can be real and edited, or generated from nothing. Detector24 asks each question with its own model, so a reviewer sees which one was answered. ### What Detector24 examines - AI-generated images: Trinity looks for images created entirely by generative models, without relying on watermarks or metadata. - Edited and forged images: Models for AI-edited image forgery, where part of a genuine photo has been changed, and for image deepfakes. - Video and audio: Video deepfake detection and face liveness, voice deepfake detection and AI-generated music detection. - Documents: Document tampering detection, document liveness, document classification and OCR for document images. - Text: Fraud text, AI-generated text and requests for personal information, along with sentiment and fake news detection. - Moderation: Image, video and text models for nudity and adult content, violence, weapons and other categories a platform needs to manage. ### Published by Detector24 - 43 specialist models in the published Detector24 catalogue. Published catalogue entries: 43. 27 image models, 6 video, 3 audio and 7 text, from AI-generated image detection and document tampering detection to voice deepfake and fraud text detection. The count of entries Detector24 publishes in its catalogue. It is not an independent validation of each model. Evidence type: Published by the portfolio entry. Source: Detector24 model catalogue. https://detector24.ai/model-catalogue (Catalogue as published on 15 October 2026). ### How Detector24 connects - Trinity in DecisionNode: Trinity is also a DecisionNode booster, so a decision can take the likelihood that an image was generated into account. - Document fraud workflows: Image-authenticity signals sit alongside document forensics in ZOREAL Verify and agentic fraud reasoning when Bynn examines the evidence behind a claim or an application. - ZOREAL Verify: ZOREAL Verify also examines images for generation and manipulation, as part of its document checks. Detector24 remains its own portfolio entry, with specialist models for every kind of media. ### Bynn offerings that now live with Detector24 These capabilities were offered on bynn.com. Detector24's specialist models carry them forward. - AI-generated image, video and audio detection: Detector24 specialist models. - Inappropriate content and media moderation: Detector24 catalogue. - Scam and fraud SMS and text moderation: Detector24 text models, including fraud text detection. Final packaging open. ## DecisionNode: AI Decision Engine and Inference URL: https://bynn.com/products/decisionnode DecisionNode is Bynn's AI decision engine: a System One decision model and inference service that turns text, images and structured inputs into decisions. ### Intelligence that makes decisions. Takes in: Text, images and structured data, with the question to answer. Returns: Typed decisions with confidence, for a system to act on. Status: Available. Available as an inference service. Benchmark figures are Bynn's own preliminary results, published with their method and the tasks where other models lead. Own site: https://decisionnode.com Pricing: From $0.035 per 1M input tokens, output free (DecisionNode Flash), as published at https://decisionnode.com/#pricing ### The part of the process that decides what happens next. DecisionNode interprets text, images and structured inputs, then returns decisions that connected systems can act on, in workflows, applications and robotics. DecisionNode is Bynn's proprietary System One decision model and inference service for workflows, applications and robotics. Specialist boosters add evidence to a decision. Sessions maintain context across a stream of inputs. From routing work to guiding a machine, DecisionNode puts AI inside the process that determines what happens next. It decides; the connected system acts. In a robot or a vehicle, DecisionNode answers the questions and the local control system moves the machine. ### Inputs in. A typed decision out. DecisionNode answers in the shape the workflow asked for, with a confidence value the system uses to decide whether to act or ask for review. ### What DecisionNode provides - Two model tiers: DecisionNode MoE for harder decisions and DecisionNode Flash for short decisions where response time matters most. - Typed decisions: Answers come back in the shape the workflow asked for, with a confidence value the system can use to decide whether to act or ask for review. - Boosters: Specialist models that add evidence to a decision: Human Age from Agemin, AI-generated image detection from Trinity, face detection and recognition, OCR, Depth Vision and Segment Vision. - Sessions: Context that persists across a stream of inputs, such as camera frames from a moving system or steps in a long workflow. - Inference API: Decisions served over an API, so an existing application can call DecisionNode at the point where it needs an answer. ### Preliminary evaluation, run by Bynn - 81.6% mean accuracy across 15 text suites in Bynn's preliminary October 2026 evaluation. Mean accuracy, 15 text suites: 81.6%. DecisionNode MoE's result in Bynn's preliminary October 2026 evaluation. The suites contain 25,134 items and cover a range of decision tasks. The published comparison includes results where other models lead. Bynn's own preliminary evaluation, run on development splits. Not an independent result. A mean across suites, not a guarantee of accuracy on any single task. Evidence type: Bynn preliminary evaluation. Source: DecisionNode benchmarks and methodology. Preliminary results, October 2026. https://decisionnode.com/benchmarks (Preliminary results, October 2026). - In Bynn's preliminary measurement, DecisionNode Flash returned a short decision in 19 ms end to end. End to end, one short decision: 19 ms. Measured from one client with the network included. About 5 ms of that time was spent on the server. One short decision on the Flash model, measured from a single client. Bynn's own preliminary measurement. Not a latency guarantee for other inputs, models or networks. Evidence type: Bynn preliminary evaluation. Source: DecisionNode benchmarks. Preliminary results, October 2026. https://decisionnode.com/benchmarks (Preliminary results, October 2026). ### How DecisionNode connects - Zora's System One: Inside Zora, DecisionNode answers the tactical questions about each camera frame. Zora plans the motion, checks every move and executes the mission. - Agemin and Trinity as boosters: Bynn's own specialist models feed DecisionNode: Agemin's age estimate and Detector24's Trinity model add evidence when a decision needs it. - Workflows that check themselves: The same approach extends to software: checking an AI agent's proposed action before it runs, or routing a case to the right next step. - Drone detection: DecisionNode is applied in drone-detection and situational-awareness systems, including systems designed to recognize groups of drones. It supplies decisions to those systems; it carries no weapons capability. ## Zora: Autonomy Engine and Mission Executive URL: https://bynn.com/products/zora Zora is Bynn's experimental autonomy engine and mission executive, linking a mission objective to perception, decisions and execution. Shown in simulation. ### The Autonomous Mind. Takes in: A mission objective, camera frames and telemetry. Returns: A mission carried through, with recovery when confidence drops. Status: Research. Zora is an experimental research platform. Zora Air demonstrates autonomous flight in simulation; Zora Ground and Zora Robotics are planned. It is not offered as a product. ### Software that sees a mission through. Zora is Bynn's mission executive and autonomy engine: it connects a mission objective with perception, decisions, planning and execution, starting with aerial missions. Zora is the intelligence behind an autonomous system, not the vehicle. Mission Control defines the mission. DecisionNode supplies tactical decisions. Zora coordinates the mission and its execution. Once a mission starts, Zora does not hand control back when something goes wrong: low confidence or a lost target triggers its own recovery. Search and rescue gives autonomous systems a clear purpose: helping people when reaching them is difficult. We are developing Zora for applications including search and rescue, firefighting and infrastructure inspection. ### From a mission objective to a mission carried through. Mission Control defines the mission. DecisionNode supplies tactical decisions. Zora coordinates the mission and its execution, and does not hand control back when something goes wrong. ### How Zora is built - Mission Control: Where an operator composes a mission as a workflow and observes it as it runs. - Mission executive: Coordinates the phases of a mission, from take-off to landing, and decides how to recover when a phase fails. - Autonomy engine: Turns the mission and what the vehicle sees into action: it tracks the target, plans the motion, checks every move and manages execution. - Zora Air (Research): Autonomous aerial missions, demonstrated in simulation with take-off, obstacle traversal and landing. - Zora Ground (Coming soon): The same engine for ground vehicles. - Zora Robotics (Coming soon): The same engine for robotic systems. ### Shown in simulation - Zora Air flies autonomous missions in simulation. Missions are composed in Mission Control and, once started, flown by the autonomy engine without handing control back: take-off, observation, obstacle traversal and landing. DecisionNode answers the tactical questions about each camera frame; Zora plans the motion, checks every move and recovers from setbacks. Simulation only. Physical vehicles and onboard computers are not yet implemented. Zora is an experimental research platform, not a product on sale. Evidence type: Simulation. Source: Zora project documentation, Bynn Intelligence. https://bynn.com/products/zora (Status as of October 2026). ### How Zora connects - DecisionNode decides, Zora executes: Tactical questions about each camera frame go to DecisionNode: where is the target, is the path clear, how risky is the next move. Its answers steer the mission. They never command the motors directly; Zora's local engine does that. - Part of Bynn's research: Zora carries Bynn's mission, advancing autonomous intelligence for the benefit of humanity, into the physical world. Its progress is reported on the research page. ## Document Fraud Detection with AI URL: https://bynn.com/document-fraud-detection AI document fraud detection for banks and insurers: tampered and fabricated documents, AI-generated images, localized forgery and deepfakes, in one workflow. ### Understand the evidence. Automate the work around it. Bynn identifies signs of altered, fabricated and AI-generated evidence, from financial documents to images submitted with a claim, and automates the work of collecting it. ### A core part of Bynn's research. Document fraud detection is where Bynn began, and it remains a core part of our AI research and portfolio. Bynn's foundations are in forensic intelligence: understanding how deception works and how to recognize it. Document forensics, image-authenticity models and agentic fraud reasoning address different parts of the problem. Together with automated collection workflows, they connect the work of obtaining evidence with the work of assessing it. ### Five different problems, not one Document fraud is a family of techniques. Each leaves different traces, so each needs its own kind of examination. - Document tampering: A genuine document with something changed: an amount, a date, a name. The page still looks right; the file usually does not. Example: A repair invoice with its total increased after it was issued. - Fabricated documents: A document that never existed, built from a template or generated from scratch to look like a real statement, payslip or certificate. Example: A bank statement produced from a template bought online. - AI-generated images: Photos created entirely by an image generator, with no camera involved, often without any watermark or metadata to reveal it. Example: A generated photo of storm damage submitted with a claim. - Localized image forgery: A real photo with one region edited, so that most of the image is authentic and only the detail that matters is not. Example: A genuine photo of a car with a dent added to one door. - Deepfakes: Faces and voices swapped or synthesized in images, video and audio, used to impersonate a real person. Example: A video of an applicant whose face has been replaced with someone else's. ### How the work is done Four capabilities, each answering a different question, working in one workflow. #### Document forensics Examines how a document was made, not only what it says. Because the evidence lies in the file's construction, the analysis works across formats and languages. - File structure, metadata and the traces left by editing software - Fonts, spacing and graphical elements that do not match across fields - Embedded digital signatures, and whether content changed after signing - Findings with the reasons behind them, for a reviewer to act on #### Image-authenticity models Specialist models that treat generation and editing as separate questions, so a reviewer knows not only that an image is suspicious but why. - Fully AI-generated images, detected from the pixels without relying on watermarks - Localized edits within an otherwise genuine photo - Face swaps in images and video, and synthesized voices in audio #### Agentic fraud reasoning AI reasoning applied to fraud assessment within a wider evidence review. An orchestrator coordinates specialist agents, either in a fixed order or chosen for the evidence in front of it, and brings their findings together. - Checks chosen for the type of evidence submitted - Findings combined into one assessment with its reasons - The plan the orchestrator follows can be reviewed and changed #### AI document collection Obtaining evidence is often slower than assessing it. Collection workflows request what is needed, follow up on what is missing and identify each document as it arrives, across languages. - Requests and reminders in the recipient's language - Each document identified and its key fields extracted - Structured output delivered to the system that needs it ### In banking and insurance #### An insurance claim Automating collection reduces administrative work. Detecting suspicious evidence addresses a different cost: losses from invalid, incorrect or fraudulent claims. - Collect: Invoices, reports and photos are requested from the claimant and followed up automatically. - Examine: Documents go through forensics; photos through image-authenticity models. - Reason: Agentic fraud reasoning weighs the findings and routes the claim, with optional human review when an agent is unsure. #### A loan application Lenders rely on statements, payslips and company records supplied by the applicant. Each can be examined before an underwriter relies on it. - Collect: Missing statements and records are requested until the application is complete. - Examine: Each document is checked for edits, reused templates and inconsistencies. - Route: Clean files go to underwriting; suspicious ones go to the fraud team with the reasons. ### Where the portfolio fits Two portfolio entries examine documents and media, each with its own focus. - ZOREAL Verify: Document and image forensics within ZOREAL, including PDFChecker's analysis of PDFs and document images, gathered into an authenticity report. Its checks run through Bynn today. Bynn's document fraud detection is moving into ZOREAL Verify. https://bynn.com/products/zoreal#verify - Detector24: Specialist models for AI-generated images, edited imagery, deepfakes and synthetic voices. https://bynn.com/products/detector24 The boundary between ZOREAL Verify and Detector24 is still being defined. Both remain available while it is. ### Bynn solutions that continue here - Insurance fraud detection: Insurance industry page. - Loan underwriting: Banking industry page. - Merchant onboarding: Banking and retail industry pages. - Tenant screening: Document fraud detection and identity verification, applied to rental applications. - Accounts-payable fraud: Document fraud detection applied to supplier invoices and payment details. - Utility-bill verification: Document forensics; packaging with identity and collection being detailed. - Bank verification: Retained; destination being confirmed. - Customer risk assessment: Retained; its relationship to agentic fraud reasoning is being defined. ### Questions about document fraud detection Q: What is document fraud detection? A: The examination of a document for signs that it was altered, fabricated or generated. Good detection looks at how the file was made as well as what it shows, because the most convincing forgeries look perfect on the page. Q: Which documents can be examined? A: PDFs and document images such as bank statements, payslips, invoices, bills, reports and certificates, along with photos submitted as evidence. Q: Can it detect AI-generated documents and images? A: Yes. Image-authenticity models look for the traces image generators leave, without relying on watermarks or metadata, and document forensics looks for the construction of a fabricated file. Q: Does it replace our fraud investigators? A: No. It examines every document so investigators can focus on the ones that need them, and it gives them specific findings rather than an unexplained score. Q: How does this relate to ZOREAL Verify? A: ZOREAL Verify is where Bynn's document forensics lives within ZOREAL, including PDFChecker. Bynn's document fraud work is moving into ZOREAL Verify, and its checks run through Bynn today. Q: Can collection and detection run in one workflow? A: Yes. Collection, forensics, image analysis and agentic fraud reasoning can run as one autonomous workflow, with human review wherever you want it. ## Enterprise AI by Industry URL: https://bynn.com/industries Enterprise AI by industry: Bynn's AI for banking, insurance, energy, retail, healthcare, telecom, public services, smart cities, public safety and defense. ### Enterprise AI, sector by sector. Document forensics, identity verification and AI workflows, configured for the documents and decisions each sector runs on. ### The same research, applied to very different work. A bank and a claims department both depend on documents they did not create. A utility and a city both run infrastructure that needs decisions made quickly and safely. Each page below starts from the work a sector actually does: the documents, the decisions, the rules and the risks. Banking and insurance are where Bynn's document work runs deepest. Other sectors are described as applications of the same capabilities, and each page says plainly how far each application has come. - Banking: Bring document forensics, identity verification and AI workflows into onboarding, lending and financial-document review. Examine the bank statements, payslips and company records behind an application, and automate the work of collecting them. https://bynn.com/industries/banking - Insurance: Automate document intake and examine claim evidence with AI. Multilingual collection, document fraud detection, image analysis and agentic fraud reasoning work together to support better claims decisions. https://bynn.com/industries/insurance - Energy and utilities: Explore AI for infrastructure inspection and the operational work around it, including autonomy research for power-line inspection. Automate the documents and decisions behind new connections, field work and customer accounts. https://bynn.com/industries/utilities - Retail: Apply image analysis, content moderation and age assurance to digital commerce and customer interactions. Check seller documents, listing images and the photos behind a refund claim. https://bynn.com/industries/retail - Healthcare: Explore identity verification, document collection and administrative automation that help teams spend less time coordinating information. The work stays administrative: patient identity, referrals, claims paperwork and staff credentials. https://bynn.com/industries/healthcare - Telecommunications: Bring identity verification, document checks and AI workflow automation to customer onboarding and service operations. Examine the scam messages and cloned voices that reach customers through your network. https://bynn.com/industries/telecommunications - Public services: Explore AI for document collection, verification and administrative workflows across languages. Reduce the manual coordination between an application and a decision, with people in charge of the outcome. https://bynn.com/industries/public-services - Smart cities: Explore decision engines for urban infrastructure, from traffic-signal decisions to automated operational workflows. DecisionNode provides decisions that connected systems can act on. https://bynn.com/industries/smart-cities - Public safety: AI for the challenges facing emergency responders, police and border-protection teams. Explore document and media authenticity analysis, evidence-review workflows and autonomy research for search and rescue. https://bynn.com/industries/public-safety - Defense: Explore DecisionNode's application in drone-detection and situational-awareness systems, including systems designed to recognize groups of drones. Decisions for detection and awareness, with no weapons capability. https://bynn.com/industries/defense Applications described on these pages are uses of Bynn's capabilities, not claims of customer deployments in every sector. ## AI for Banking: Onboarding, Lending and Document Fraud URL: https://bynn.com/industries/banking Bynn brings document forensics, identity verification and AI workflows to onboarding, lending and financial-document review in banking and fintech. ### Know what stands behind an application. Bring document forensics, identity verification and AI workflows into onboarding and financial-document review. Assess the evidence behind an application and automate the work needed to collect it. ### Every application rests on documents the bank did not create. Opening an account, approving a loan or onboarding a merchant depends on evidence supplied by the applicant: identity documents, bank statements, payslips, tax returns, certificates of incorporation, proof of address. Each one is a point where an altered figure or a fabricated record can enter. Generative tools have made convincing statements and payslips cheap to produce, and synthetic identities combine real and invented details. Review teams are asked to move faster while the documents in front of them become harder to trust. ### What banking teams answer to - Know your customer and anti-money laundering: Onboarding obligations require evidence about customers and businesses that holds up when an auditor or supervisor asks how a decision was made. - High-risk AI in lending: The EU AI Act classes AI used to assess the creditworthiness of individuals as high-risk, which raises the bar for documentation, oversight and explainable findings. - Impersonation at a distance: Deepfake video and cloned voices are used to impersonate customers and executives in the calls where payments and changes are approved. ### Where AI does the work - Loan and mortgage document review (Uses available products): Bank statements, payslips, tax returns and collateral documents are examined for tampering and fabrication before an underwriter relies on them. - Identity verification at onboarding (Uses available products): Document and biometric identity checks confirm the applicant is the person on the document and is present, through ZOREAL's enterprise identity verification. - Merchant and business onboarding (Uses available products): Certificates of incorporation, business licences, tax documents and bank letters are collected and examined as part of know-your-business review. - Automated document collection (Uses available products): Agents request missing documents, follow up and identify each file as it arrives, so an application is complete before review begins. - Verified calls for high-risk requests (Application direction): ZOREAL Meet lets a customer or colleague prove they are a verified, live person before a payment instruction or account change is accepted. ### A small-business loan, end to end One example of how the pieces fit. Each step can run on its own or route to a person, depending on your policy. - Collect: The applicant is asked for identity documents, twelve months of bank statements and the latest accounts. Missing items are chased automatically. - Verify identity: The directors' identity documents are checked and matched to a live capture. - Examine documents: Statements and accounts are examined for edits, reused templates and inconsistencies in structure, fonts and metadata. - Reason and route: Agentic fraud reasoning weighs the findings. A clean file goes to underwriting; a suspicious one goes to the fraud team with the reasons. - Decide: The underwriter decides on evidence that has already been checked, and the findings stay on file for review. ### What is involved - ZOREAL: Enterprise identity verification and verified meetings. https://bynn.com/products/zoreal - Document fraud detection: Forensic examination of financial documents. https://bynn.com/document-fraud-detection - ZOREAL Verify: PDF and document image forensics. https://bynn.com/products/zoreal#verify - AI document collection: Requests and follow-up for application documents. https://bynn.com/document-fraud-detection#document-collection - DecisionNode: Decisions that route each case to its next step. https://bynn.com/products/decisionnode Document examination, identity verification and collection use available offerings. Verified calls for payment approvals are an application of ZOREAL Meet, not a described deployment. ### Questions from banking teams Q: Which documents can you examine? A: The documents lending and onboarding depend on, including bank statements, payslips, tax returns, invoices, company records and identity documents, as PDFs or images. Q: Does this make the credit decision? A: No. It examines the evidence and routes cases. The credit decision stays with your underwriters and your models, with the findings available to support it. Q: How does this relate to ZOREAL? A: Bynn's enterprise identity verification is moving to ZOREAL Verify. A banking engagement can use it alongside Bynn's document forensics and workflow automation. ## Insurance Claims Automation and Fraud Detection URL: https://bynn.com/industries/insurance Better claims decisions start with better evidence. Bynn automates claims document collection across languages and examines claim evidence for fraud. ### Better claims decisions start with better evidence. Bynn automates the work of collecting claim documents across languages, then examines the evidence for signs of fabrication or manipulation before a decision is made. ### Most of a claim's time is spent waiting for evidence. A claim can generate repeated requests and long email exchanges before anyone can assess it. Repair invoices, medical bills, police reports and photos of damage arrive in different formats and languages, often incomplete. Some of that evidence may not be what it seems: an invoice with a changed amount, a reused template, a damage photo that was generated or edited. Staged or exaggerated, fraud drives up costs for honest policyholders. ### What claims teams answer to - Fair, prompt handling: Policyholders and regulators expect valid claims to be settled quickly and fairly. Slow evidence gathering works against both. - Investigations that hold up: When special investigation units review a suspicious claim, they need specific findings they can explain, not an unexplained score. - Generated evidence: Image generators and editing tools make convincing damage photos and documents easy to produce, and hard to spot by eye. ### Where AI does the work - Automated document collection (Uses available products): Agents request the documents a claim needs, follow up on what is missing and handle the correspondence across languages, so the evidence reaches the claims process without manual chasing. - Document fraud detection (Uses available products): Invoices, bills and reports are examined for tampering, reused templates and hidden edits, using the file's structure as well as its appearance. - Image authenticity (Uses available products): Photos of damage are checked for signs of AI generation and localized editing, as separate signals. - Agentic fraud reasoning (Uses available products): AI reasoning brings the findings together within the wider evidence review, so a suspicious claim reaches an investigator with the reasons attached. ### One claim, from first notice to decision Run agentic workflows autonomously, with optional human review when an agent is unsure how to proceed. Configure the workflow around the way your organization handles claims. - Request: A claim is opened. The workflow works out which documents are needed and asks the claimant for them, in the claimant's language. - Follow up: Missing or unreadable documents trigger a follow-up instead of an email thread. Each document is identified as it arrives. - Examine: Document forensics and image-authenticity models examine the evidence for indicators of fabrication or manipulation. - Reason: Agentic fraud reasoning weighs the findings in context. Clear claims move on; uncertain ones are routed to a person with the evidence and the reasons. - Decide: The claims handler or investigator makes the decision with complete, examined evidence in front of them. ### Two benefits, measured separately - Administrative time saved: Automating collection removes the repeated requests and correspondence needed to obtain the right documents. This is a saving in handling time. - Fraud losses avoided: Detecting suspicious evidence addresses a different cost: losses from invalid, incorrect or fraudulent claims. This is a saving in what is paid out. ### What is involved - AI document collection: Requests, follow-up and intake across languages. https://bynn.com/document-fraud-detection#document-collection - Document fraud detection: Forensic examination of submitted documents. https://bynn.com/document-fraud-detection#document-fraud - Agentic fraud reasoning: Reasoning across findings within the review. https://bynn.com/document-fraud-detection#fraud-reasoning - Detector24: Specialist models for AI-generated and edited images. https://bynn.com/products/detector24 - ZOREAL Verify: Forensic analysis of PDFs and document images. https://bynn.com/products/zoreal#verify Claims document collection and evidence examination are available to insurers today. Bynn does not publish customer names or savings figures until they can be stated with their scope and evidence. ### Questions from claims teams Q: Does this replace claims handlers? A: No. It removes the coordination work around a claim and examines the evidence. Decisions on uncertain or suspicious claims go to your people, with the findings attached. Q: Can the workflow run without human review? A: Yes, if you configure it that way. Workflows run autonomously by default, and you choose where a person reviews, including whenever an agent is unsure how to proceed. Q: Which documents can be examined? A: The documents claims depend on, such as invoices, bills, reports and statements as PDFs or images, along with photos submitted as evidence. Q: How much will we save? A: It depends on your claim volumes, lines of business and current process. We measure administrative time and fraud losses separately with you rather than quote a generic figure. ## AI for Energy and Utilities: Inspection and Operational Workflows URL: https://bynn.com/industries/utilities Enterprise AI for energy and utilities: autonomy research for power-line inspection, decisions on inspection imagery and automated customer and field work. ### Infrastructure that is inspected, understood and acted on. Explore AI for infrastructure inspection and operational workflows, including autonomy research for power-line inspection. ### Long networks, small teams, and a lot of paperwork. Power, water and gas networks run across distances no inspection team can cover often enough. Vegetation grows into lines, equipment corrodes and storms leave damage that has to be found before it can be fixed. Most of the imagery collected is never looked at closely. On the customer side, the work is administrative: new connection applications, changes of occupancy, evidence for social tariffs and hardship support, and identity checks before an account is changed. Much of it still moves by email and attachment. ### What energy and utility operators answer to - Critical infrastructure rules: Energy and water operators fall under critical-infrastructure security rules such as the EU's NIS2 directive, and AI used as a safety component in their networks is classed as high-risk under the EU AI Act. - Safe inspection: Inspecting lines and assets is slow and can be dangerous. Any autonomous system near live infrastructure has to be evaluated before it flies. - Account fraud: Changes of name, address or bank details are a route to account takeover and fraudulent claims for support. ### Where AI does the work - Autonomous power-line inspection (Research): Zora is being developed for infrastructure inspection: a mission executive that coordinates an inspection flight and responds to what it observes. Demonstrated in simulation only. - Decisions on inspection imagery (Application direction): DecisionNode can classify what an inspection image shows, such as vegetation encroachment or visible damage, and route it to the right maintenance queue. - Customer documents, collected and checked (Uses available products): Connection applications and support claims gather their documents automatically, and submitted bills, statements and proofs of address are examined for tampering. - Verified account changes (Uses available products): Identity verification confirms the person requesting a change of account holder or bank details is who they claim to be. ### From an inspection image to a work order An application direction that shows how a decision engine fits the work. It describes how the pieces could connect, not a running deployment. - Observe: Inspection imagery arrives from a patrol, a fixed camera or, in future, an autonomous mission. - Decide: DecisionNode classifies each image and returns a typed decision with its confidence: no action, monitor, or repair. - Route: Repair decisions become work orders in the maintenance system; low-confidence images go to an engineer. - Act: Field teams receive prioritized work with the image and the reason attached. ### What is involved - Zora: Autonomy research for inspection missions, in simulation. https://bynn.com/products/zora - DecisionNode: Decisions on imagery and operational inputs. https://bynn.com/products/decisionnode - AI document collection: Customer document requests and follow-up. https://bynn.com/document-fraud-detection#document-collection - ZOREAL: Identity verification for account changes. https://bynn.com/products/zoreal Inspection autonomy is research shown in simulation. Imagery decisions are an application direction. Customer document and identity workflows use available offerings. ### Questions from energy and utility teams Q: Is the inspection autonomy available today? A: No. Zora's inspection work is research, demonstrated in simulation only. The customer document and identity workflows on this page use offerings that are available today. Q: Does DecisionNode control field equipment? A: No. It returns a typed decision with its confidence, such as no action, monitor or repair. Your maintenance systems and engineers act on it, and low-confidence images go to a person. Q: Which customer documents can be checked? A: Bills, statements and proofs of address submitted with connection applications, support claims and account changes are examined for tampering. Identity verification confirms who is asking to change an account holder or bank details. ## AI for Retail and E-commerce: Trust and Automation URL: https://bynn.com/industries/retail Enterprise AI for digital commerce: seller verification, AI-generated image detection, content moderation, age assurance and automated customer workflows. ### Commerce runs on images, reviews and trust. Apply image analysis, content moderation, age assurance and automated workflows to digital commerce and customer interactions. ### Every listing, review and refund is a claim about the world. A marketplace depends on what sellers say about themselves and their products, and what buyers say back. Product photos, seller documents, reviews and the photos sent with a refund request all shape decisions that happen at scale. Generated images make it cheap to flood a catalogue with listings that do not exist, or to fake damage on a delivered item. Age-restricted products need checks that do not drive honest customers away at checkout. ### What retail teams answer to - Knowing your traders: Under the EU Digital Services Act, online marketplaces must collect and check information about the traders who sell on them. - Age-restricted sales: Alcohol, tobacco, knives and other restricted goods need an age check online that is proportionate to the product. - Refund and listing abuse: Edited or generated photos support false refund claims and fake listings, and the cost lands on margins and on buyer trust. ### Where AI does the work - Seller onboarding (Uses available products): Business registrations, tax documents and bank letters are collected and examined, and the people behind a seller account are identity-checked. - Refund evidence (Application direction): Photos submitted with a damage or non-delivery claim are checked for AI generation and localized editing before a refund is approved. - Listing and review moderation (Uses available products): Detector24 models flag AI-generated product images and moderate images and text against the categories a marketplace prohibits. - Age assurance at checkout (Uses available products): Agemin estimates age from a face or an email address for age-restricted products, with an ID check when a higher level of certainty is required. ### A damaged-item refund claim An application of available models to a common retail process. It describes how the steps connect, not a customer deployment. - Submit: A customer reports a damaged item and uploads photos. - Examine: Each photo is checked for signs of AI generation and of localized edits, as separate signals. - Decide: DecisionNode combines the image signals with order details and returns a decision: refund, request more evidence, or review. - Act: Clear claims are refunded at once. Uncertain ones go to an agent with the findings, so honest customers are not delayed. ### What is involved - Detector24: AI-generated image detection and moderation models. https://bynn.com/products/detector24 - Agemin: Age assurance for restricted products. https://bynn.com/products/agemin - ZOREAL: Identity verification for sellers. https://bynn.com/products/zoreal - DecisionNode: Decisions that route claims and listings. https://bynn.com/products/decisionnode Moderation, age assurance and seller verification use available offerings. Refund-evidence checks are an application of the same models, not a described deployment. ### Questions from retail and marketplace teams Q: Can a refund be decided automatically? A: In the workflow on this page, clear claims are refunded at once and uncertain ones go to an agent with the findings. Refund-evidence checks are an application of available models, not a described deployment. Q: Are AI-generated and edited photos treated the same? A: No. Each photo is checked for signs of AI generation and for localized edits as separate signals, so a reviewer can see which one was found. Q: How does age assurance work at checkout? A: Agemin estimates age from a face or an email address. When a product needs a higher level of certainty, an ID check can follow. ## AI for Healthcare Administration and Patient Identity URL: https://bynn.com/industries/healthcare Enterprise AI for healthcare administration: patient identity verification, document collection, claims paperwork and staff credential checks. No diagnostics. ### Less time coordinating. More time for patients. Explore identity verification, document collection and administrative automation that help teams spend less time coordinating information. ### Behind every appointment is a stack of administration. Registering a patient, starting a telehealth consultation, processing a referral or submitting a claim all depend on identity and paperwork. Staff spend hours chasing forms, insurance details and records from other providers. Errors in identity are costly. Medical identity theft puts the wrong history on a patient's record, and forged credentials put unqualified people into roles that require a licence. This is administrative work, and it is where Bynn's capabilities apply. ### What healthcare administrators answer to - Health data protection: Health information is special-category data under GDPR and protected health information under HIPAA in the United States. Processing has to be limited to what the purpose needs. - The right patient: Remote care makes it harder to confirm that the person on the screen is the patient whose record is open. - Verified staff: Clinical roles depend on diplomas, licences and registrations that can be forged. ### Where AI does the work - Patient identity at registration and telehealth (Uses available products): Document and biometric identity checks confirm a patient's identity before a record is created or a remote consultation begins. - Referral and intake documents (Uses available products): Agents request the forms, insurance details and records an intake needs, follow up on what is missing and identify each document as it arrives. - Claims paperwork (Uses available products): Medical invoices and supporting documents submitted for reimbursement are examined for alteration before payment. - Staff credential checks (Uses available products): Diplomas, licences and certificates supplied during hiring are examined for tampering and fabrication. ### A new patient, ready before the first appointment An administrative workflow built from available capabilities. Clinical decisions are not part of it. - Invite: The patient receives a request for identity and insurance documents in their own language. - Verify: Their identity document is checked and matched to a live capture. - Collect: Referral letters and previous records are requested and followed up automatically. - Hand over: The care team receives a complete, verified intake, and staff are only involved when something needs judgement. ### What is involved - ZOREAL: Patient identity verification and verified video calls. https://bynn.com/products/zoreal - AI document collection: Intake and referral documents. https://bynn.com/document-fraud-detection#document-collection - Document fraud detection: Invoices and credential documents. https://bynn.com/document-fraud-detection These are applications of available identity, collection and document capabilities to healthcare administration. Bynn offers no diagnostic or clinical capability. ### Questions from healthcare administrators Q: Does Bynn make clinical or diagnostic decisions? A: No. Bynn offers no diagnostic or clinical capability. The work on this page is administrative: patient identity, intake and referral documents, claims paperwork and staff credentials. Q: Can patients verify their identity remotely? A: Yes. Document and biometric identity checks through ZOREAL confirm a patient's identity before a record is created or a remote consultation begins. Q: When are staff involved? A: When something needs judgement. Documents are requested and followed up automatically, and the care team receives a complete, verified intake. ## AI for Telecommunications: Onboarding and Fraud URL: https://bynn.com/industries/telecommunications Enterprise AI for telecom operators: identity checks at SIM registration, device-financing document review, scam message detection and service automation. ### Trust on the network starts at the counter. Bring identity verification, document checks and AI workflow automation to customer onboarding and service operations. ### The network is how fraud reaches people. Operators verify identities when a SIM is registered, a contract is signed or a device is financed. Weak checks lead to SIM-swap fraud, fraudulent porting and contracts taken out in someone else's name. The same network carries the scam texts and voice-cloned calls aimed at an operator's own customers. A few seconds of audio is now enough to clone a voice, so a call that sounds right is no longer proof of who is speaking. ### What operators answer to - SIM registration rules: Many countries require operators to verify a subscriber's identity before activating a SIM. - Account takeover: SIM swaps and number ports are used to intercept one-time codes and take over bank and email accounts. - Scam traffic: Operators are under growing pressure to detect and stop scam messages before they reach customers. ### Where AI does the work - Identity checks at registration and SIM swap (Uses available products): Document and biometric identity checks confirm the subscriber before a SIM is activated or moved to a new card. - Device-financing documents (Uses available products): Payslips and bank statements supplied for a financed handset are examined for tampering. - Scam message detection (Application direction): Detector24's fraud text models classify messages for scam and phishing content, as part of a wider filtering policy. - Voice clone detection (Application direction): Voice deepfake detection examines audio for signs of synthesis, and ZOREAL Meet anchors identity on face liveness rather than voice. ### A SIM-swap request How available capabilities could protect one of the riskiest requests an operator handles. - Request: A customer asks for their number to be moved to a new SIM, in store or online. - Verify: Their identity document is checked and matched to a live capture, instead of answers to security questions. - Decide: The request proceeds, is held for a cooling-off period, or goes to the fraud team, according to policy. - Record: The verification result is kept with the request for any later dispute. ### What is involved - ZOREAL: Identity verification and liveness. https://bynn.com/products/zoreal - Detector24: Fraud text and voice deepfake models. https://bynn.com/products/detector24 - Document fraud detection: Documents supplied for contracts and financing. https://bynn.com/document-fraud-detection - DecisionNode: Decisions in service workflows. https://bynn.com/products/decisionnode Identity and document checks use available offerings. Network-scale scam and voice-clone detection are application directions for the same models, not described deployments. ### Questions from operators Q: Can identity be checked before a SIM swap? A: Yes. Document and biometric identity checks confirm the subscriber before a SIM is activated or moved to a new card. Q: Can scam messages be detected on our network? A: Detector24's fraud text models classify messages for scam and phishing content as part of a wider filtering policy. Network-scale detection is an application direction for these models, not a described deployment. Q: Why not verify customers by voice? A: Cloned voices make voice a weak factor. Voice deepfake detection examines audio for signs of synthesis, and ZOREAL Meet anchors identity on face liveness rather than voice. ## AI for Public Services: Documents and Casework URL: https://bynn.com/industries/public-services Enterprise AI for public organizations: multilingual document collection, verification and administrative workflows, with people in charge of every outcome. ### Services that move at the speed people need. Explore AI for document collection, verification and administrative workflows, helping public organizations reduce manual coordination and deliver services more efficiently. ### Between an application and a decision, there is a lot of chasing. Benefits, permits, licences and registrations all start with an application and a set of supporting documents. Caseworkers spend much of their time requesting what is missing, explaining requirements and checking what arrives, often across several languages. Forged and altered documents enter the same queues, and the pressure to clear backlogs makes them harder to catch. ### What public organizations answer to - High-risk decisions about people: The EU AI Act classes AI used to assess eligibility for public benefits and services as high-risk. People must stay in charge of those decisions. - Equal access: Services must work for residents who speak different languages and have different levels of digital confidence. - Accountability: Every step of a case may need to be explained to the applicant, an auditor or a court. ### Where AI does the work - Multilingual document collection (Uses available products): Agents request the documents an application needs and follow up in the applicant's language, so caseworkers start with a complete file. - Document verification (Uses available products): Supporting documents are examined for tampering and fabrication, and each finding comes with the reason it was raised. - Identity verification for digital services (Uses available products): Document and biometric identity checks let residents prove who they are online, without an office visit. - Case routing (Application direction): DecisionNode can sort incoming cases by type and completeness so they reach the right team, while eligibility decisions stay with caseworkers. ### A permit application How available capabilities connect in a typical administrative process. The decision itself stays with a caseworker. - Apply: A resident submits an application online and proves their identity. - Complete: The workflow requests the missing documents in the resident's language and follows up until the file is complete. - Check: Documents are examined for tampering, and anything unusual is flagged with its reason. - Decide: A caseworker reviews a complete, checked file and makes the decision. ### What is involved - AI document collection: Multilingual requests and follow-up. https://bynn.com/document-fraud-detection#document-collection - Document fraud detection: Examination of supporting documents. https://bynn.com/document-fraud-detection - ZOREAL: Identity verification for online services. https://bynn.com/products/zoreal - DecisionNode: Routing decisions inside casework. https://bynn.com/products/decisionnode These are applications of available collection, verification and identity capabilities to public administration. No government deployment is described. ### Questions from public organizations Q: Who makes eligibility decisions? A: Caseworkers. DecisionNode can sort incoming cases by type and completeness so they reach the right team, and eligibility decisions stay with your people. Q: Can applicants answer in their own language? A: Yes. Agents request the documents an application needs and follow up in the applicant's language, so caseworkers start with a complete file. Q: Is this in use in government today? A: These are applications of available collection, verification and identity capabilities to public administration. No government deployment is described. ## AI for Smart Cities: Decision Engines for Infrastructure URL: https://bynn.com/industries/smart-cities How an AI decision engine can support urban infrastructure, from traffic-signal decisions to operational workflows. An application direction for DecisionNode. ### Decisions for the systems a city runs on. Explore decision engines for urban infrastructure, from traffic-signal decisions to automated operational workflows. DecisionNode provides decisions that connected systems can act on. ### A city makes millions of small decisions a day. How long a signal stays green, which pothole report is urgent, where a maintenance crew goes next. Most of these decisions follow fixed timetables and queues that cannot see what is happening on the street right now. Cities already collect the inputs: camera feeds, sensor data, reports with photos. A decision engine turns those inputs into choices that existing control and workflow systems can carry out. ### What cities answer to - Safety-critical infrastructure: AI used as a safety component in road traffic management is classed as high-risk under the EU AI Act. Any such system needs integration, evaluation and oversight before it acts. - Privacy in public space: Cameras in public places capture people who never agreed to be analyzed. Processing should be limited, and faces redacted where they are not needed. - Public accountability: Residents expect to understand why the systems around them behave as they do. ### Where a decision engine fits - Traffic-signal decisions (Application direction): DecisionNode can interpret camera frames at a junction and return a timing decision for the signal controller to apply within its safety limits. - Report triage (Application direction): Photos and descriptions submitted by residents can be classified and routed to the right department with a priority. - Privacy-preserving imagery (Application direction): Detector24's catalogue includes face redaction, vehicle detection and people counting, so imagery can be used for operations without exposing who is in it. - Operational workflows (Application direction): Decisions trigger the next step in maintenance and service systems, with a person reviewing anything uncertain. ### A junction that sees its own traffic An application direction, described to show the role of a decision engine. It is not a deployed or certified traffic-control system. - Observe: A camera at the junction streams frames; DecisionNode sessions keep the context from one frame to the next. - Decide: DecisionNode estimates queue lengths and returns a typed timing decision with its confidence. - Constrain: The signal controller applies the decision only within its own safety rules and fixed limits. - Review: Engineers monitor the decisions and their effect before widening their use. ### What is involved - DecisionNode: Decisions from camera frames and operational inputs. https://bynn.com/products/decisionnode - Detector24: Face redaction, vehicle detection and people counting models. https://bynn.com/products/detector24 Smart-city uses are application directions for DecisionNode and Detector24. No municipal deployment or certified traffic-control system is claimed. ### Questions from city teams Q: Does DecisionNode control traffic signals? A: It returns a timing decision, and the signal controller applies it within its own safety limits. Any use at a junction needs integration and evaluation first, and no certified traffic-control system is claimed. Q: Can city imagery be used without exposing people? A: Detector24's catalogue includes face redaction, vehicle detection and people counting, so imagery can support operations without revealing who is in it. Q: Is any of this running in a city today? A: No municipal deployment is described. Smart-city uses are application directions for DecisionNode and Detector24. ## AI for Public Safety: Evidence, Documents and Rescue URL: https://bynn.com/industries/public-safety AI for emergency responders, police and border protection: document and media authenticity, evidence-review workflows and search and rescue autonomy research. ### Evidence you can trust, under pressure. AI for the challenges facing emergency responders, police and border-protection teams. ### Decisions made under pressure, on evidence that may be fake. Investigators receive photos, video, audio and documents from the public, from devices and from other agencies. Some of it has been edited or generated, and a convincing deepfake can mislead an investigation or a court. Border teams examine travel documents at speed. Emergency responders search difficult terrain where every minute matters. Each of these is a place where better evidence, or another pair of eyes, changes the outcome. ### What public-safety teams answer to - High-risk use: The EU AI Act classes many uses of AI in law enforcement and border control as high-risk, with strict requirements for oversight and documentation. - Chain of evidence: Findings about a piece of media have to be explainable and reproducible if they are to be used. - Responder safety: Search and rescue puts teams in dangerous terrain and conditions. ### Where AI does the work - Media authenticity analysis (Uses available products): Detector24 models examine images, video and audio for signs of generation and manipulation, giving investigators distinct signals for each kind of evidence. - Document authenticity (Uses available products): Document forensics examines submitted PDFs and images, and chip-based checks read the security chip in passports and identity cards. - Evidence-review workflows (Application direction): Agentic workflows gather the submissions in a case, run the relevant checks and present the findings to an investigator with their reasons. - Search and rescue autonomy (Research): Zora is being developed to coordinate aerial search missions, responding to what it observes and managing recovery when a plan fails. Shown in simulation. ### A video submitted as evidence How available models could support an investigator. It describes the steps, not an agency deployment. - Receive: A video and its accompanying statement are added to a case. - Examine: The video is checked for deepfake signs and the audio for voice synthesis, as separate findings. - Present: The investigator sees what was checked, what was flagged and why. - Decide: The investigator decides how to treat the evidence. The analysis informs that decision; it does not make it. ### What is involved - Detector24: Image, video and audio authenticity models. https://bynn.com/products/detector24 - ZOREAL Verify: Document forensics. https://bynn.com/products/zoreal#verify - ZOREAL: Identity verification and chip-based document checks. https://bynn.com/products/zoreal - Zora: Autonomy research for search and rescue, in simulation. https://bynn.com/products/zora Media and document authenticity use available models. Evidence workflows are an application direction, and search and rescue autonomy is research shown in simulation. No agency deployment or operational outcome is described. ### Questions from public-safety teams Q: What can be checked for manipulation? A: Images, video and audio, each with its own signals for generation and manipulation, as well as submitted PDFs and document images. Chip-based checks read the security chip in passports and identity cards. Q: Who decides what the evidence means? A: The investigator. In the evidence-review workflow, an application direction, agents gather the submissions in a case, run the relevant checks and present the findings with their reasons. Q: Is the search and rescue autonomy operational? A: No. Zora's search and rescue work is research, shown in simulation. ## AI for Defense: Drone Detection and Situational Awareness URL: https://bynn.com/industries/defense AI for defense: DecisionNode in drone-detection and situational-awareness systems, including recognizing groups of drones, with no weapons capability. ### See the threat sooner. Understand it faster. Explore DecisionNode's application in drone-detection and situational-awareness systems, including systems designed to recognize groups of drones. ### Small drones have changed what has to be watched. Inexpensive drones can approach bases, borders and critical infrastructure low and fast, alone or in groups. Operators watching sensor feeds have seconds to tell a drone from a bird, and one drone from a coordinated group. The information environment has changed too. Generated video and audio are used to deceive, and analysts need to know what is authentic before they act on it. ### What defense work demands - Speed with oversight: Detection has to keep pace with the threat while people remain responsible for every response. - Groups, not single objects: Recognizing that several contacts form one coordinated group changes how a threat is understood. - Authentic information: Decisions depend on media and reports that may have been fabricated. ### Where AI does the work - Drone detection (Application direction): DecisionNode interprets sensor and camera inputs within detection systems and returns decisions about what has been observed. - Drone-group recognition (Application direction): Sessions keep context across a stream of inputs, so a system can recognize several contacts as a group rather than as separate objects. - Media authenticity for analysts (Uses available products): Detector24 models examine video, images and audio for generation and manipulation before they inform an assessment. - Reconnaissance and humanitarian survey (Research): Longer term, our autonomy research looks toward reconnaissance and helping identify mines and unexploded ordnance after conflict. This is an ambition, not a current capability. ### From a sensor contact to an operator's screen How DecisionNode fits within a detection system. It supplies decisions to the system; people decide on any response. - Sense: Cameras and sensors in the detection system report a contact. - Interpret: DecisionNode classifies the contact and, across a session, whether it is part of a group. - Inform: The system presents the decision and its confidence to the operator. - Respond: Operators decide what happens next. DecisionNode has no weapons capability and controls no effector. ### What is involved - DecisionNode: Decisions within detection and awareness systems. https://bynn.com/products/decisionnode - Detector24: Authenticity analysis for media. https://bynn.com/products/detector24 - Zora: Autonomy research, in simulation. https://bynn.com/products/zora DecisionNode's role in drone detection is described at a high level. No customer, operation or performance result is claimed, and no weapons capability is offered. ### Questions about defense applications Q: Does Bynn offer weapons capability? A: No. DecisionNode supplies decisions about what sensors and cameras have observed, within detection and situational-awareness systems. No weapons capability is offered. Q: What does drone-group recognition mean? A: Sessions keep context across a stream of inputs, so a system can recognize several contacts as one coordinated group rather than as separate objects. Q: Is this deployed with a named customer? A: DecisionNode's role in drone detection is described at a high level. No customer, operation or performance result is claimed. ## About Us: Frontier AI Research Lab URL: https://bynn.com/company About Bynn Intelligence, a frontier AI research lab building proprietary models and autonomous systems, toward a world where AI protects people. ### We build AI that protects people. Bynn Intelligence is a frontier AI research lab developing proprietary models and autonomous systems for the real world. ### Who we are Our work spans digital trust, forensics, age assurance, decision intelligence and autonomy. We are the company behind ZOREAL, Agemin, Detector24, DecisionNode and Zora. Our portfolio brings research into practical use, from assessing the authenticity of digital evidence to making decisions within workflows and coordinating autonomous missions. We also work directly with enterprises, taking AI from a business challenge through research, development, integration and the infrastructure needed to run it. Our vision: A world where AI protects people and expands what they can achieve. Our mission: Advance autonomous intelligence for the benefit of humanity. We develop decision engines and autonomous systems that can interpret complex situations, choose what to do and carry out tasks in the digital and physical world. From fraud prevention and digital trust to public safety, search and rescue and defense, our research explores how intelligent systems can protect people and take on demanding work. ### From understanding threats to building intelligence. Bynn's foundations are in fraud detection and forensic intelligence: understanding how deception works and how to recognize it. Building our own AI models and systems has expanded both our expertise and our ambition. Today, that work reaches across digital trust, age assurance, decision engines and autonomous systems. We bring it into the world through our portfolio and through the solutions we develop with enterprises. Our experience continues to shape how we build: investigate deeply, develop the intelligence, and put it to work on real problems. Our ambition is to use AI to protect people, take on complex tasks and expand what they can achieve. ### How we work - Investigate deeply: We start by understanding how a threat, a process or a decision really works, before we decide what to build. - Develop the intelligence: We build our own models, decision engines and autonomous systems, so we understand them from the inside and can adapt them to the problem. - Put it to work: Research earns its place when it runs in the world, through our portfolio and through the systems we build with enterprises. - Say exactly what is proven: Independent results, our own evaluations and simulations are different kinds of evidence. We label each one, and we do not claim what we have not shown. ### The Bynn portfolio Five portfolio entries under one lab, each with its own focus. - ZOREAL (Digital trust): Human trust for digital interactions: identity verification and document forensics in ZOREAL Verify, verified meetings, sign-in and signing. https://bynn.com/products/zoreal - Agemin (Age assurance): AI for age assurance, built on Bynn's facial age-estimation research. https://bynn.com/products/agemin - Detector24 (Media forensics): Specialist models that examine what AI can create and conceal in images, video, audio and text. https://bynn.com/products/detector24 - DecisionNode (Decision intelligence): A decision model and inference service that returns decisions connected systems can act on. https://bynn.com/products/decisionnode - Zora (Autonomy): A mission executive and autonomy engine for autonomous systems, demonstrated in simulation. https://bynn.com/products/zora ## Research and Evidence URL: https://bynn.com/company/research Bynn's AI research results by evidence type: NIST's independent age-estimation evaluation, DecisionNode's preliminary benchmarks and Zora in simulation. ### Research, with the evidence labelled. What we have measured, who measured it and under what conditions. Independent results, our own evaluations and simulations are kept apart. ### Different evidence deserves different weight. A result measured by NIST is not the same kind of evidence as a benchmark we ran ourselves, and neither is the same as a mission flown in simulation. Each result below carries its type, its source and its date, so you can weigh it for what it is. ### How to read the evidence - Independent evaluation: Measured by an independent organization under its own protocol. - Bynn preliminary evaluation: Run by Bynn, published with its method and labelled preliminary. - Simulation: Demonstrated in a simulated environment, not in the physical world. - Published by the portfolio entry: A figure a portfolio entry publishes about itself, such as a catalogue count. ### Results - #1 for true-positive rate in NIST's Child Online Safety evaluation for ages 13-16. True-positive rate, ages 13-16: 72.53%. Bynn's facial age-estimation model, which powers Agemin, achieved a 72.53% true-positive rate: the proportion of people aged 13 to 16 correctly estimated within that age band. Measured by NIST in its ongoing evaluation program. NIST evaluation is not certification or endorsement. The result measures age estimation within the 13 to 16 band. It does not measure a reduction in online harm. Evidence type: Independent evaluation. Source: NIST FATE Age Estimation & Verification. Submission bynn-000. https://pages.nist.gov/frvt/html/frvt_age_estimation.html (NIST results table last updated 29 September 2026). - 81.6% mean accuracy across 15 text suites in Bynn's preliminary October 2026 evaluation. Mean accuracy, 15 text suites: 81.6%. DecisionNode MoE's result in Bynn's preliminary October 2026 evaluation. The suites contain 25,134 items and cover a range of decision tasks. The published comparison includes results where other models lead. Bynn's own preliminary evaluation, run on development splits. Not an independent result. A mean across suites, not a guarantee of accuracy on any single task. Evidence type: Bynn preliminary evaluation. Source: DecisionNode benchmarks and methodology. Preliminary results, October 2026. https://decisionnode.com/benchmarks (Preliminary results, October 2026). - In Bynn's preliminary measurement, DecisionNode Flash returned a short decision in 19 ms end to end. End to end, one short decision: 19 ms. Measured from one client with the network included. About 5 ms of that time was spent on the server. One short decision on the Flash model, measured from a single client. Bynn's own preliminary measurement. Not a latency guarantee for other inputs, models or networks. Evidence type: Bynn preliminary evaluation. Source: DecisionNode benchmarks. Preliminary results, October 2026. https://decisionnode.com/benchmarks (Preliminary results, October 2026). - Zora Air flies autonomous missions in simulation. Missions are composed in Mission Control and, once started, flown by the autonomy engine without handing control back: take-off, observation, obstacle traversal and landing. DecisionNode answers the tactical questions about each camera frame; Zora plans the motion, checks every move and recovers from setbacks. Simulation only. Physical vehicles and onboard computers are not yet implemented. Zora is an experimental research platform, not a product on sale. Evidence type: Simulation. Source: Zora project documentation, Bynn Intelligence. https://bynn.com/products/zora (Status as of October 2026). - 43 specialist models in the published Detector24 catalogue. Published catalogue entries: 43. 27 image models, 6 video, 3 audio and 7 text, from AI-generated image detection and document tampering detection to voice deepfake and fraud text detection. The count of entries Detector24 publishes in its catalogue. It is not an independent validation of each model. Evidence type: Published by the portfolio entry. Source: Detector24 model catalogue. https://detector24.ai/model-catalogue (Catalogue as published on 15 October 2026). - [Read the research articles](https://bynn.com/tags/research) ### Where the research is going Research directions are ambitions with work behind them. Their status is stated plainly. - Facial age estimation (Available): Improving age estimation for the age bands that matter most for online safety, evaluated through NIST's ongoing program. - Decision models (Available): DecisionNode's System One models for workflows, applications and robotics, with boosters and sessions that add evidence and context. - Autonomy for search and rescue (Research): Zora's mission executive and autonomy engine, demonstrated in simulation, developed toward search and rescue, firefighting and infrastructure inspection. - Ground and robotic autonomy (Coming soon): Zora Ground and Zora Robotics share the same engine and are planned. - Authenticity of evidence (Available): Forensic and image-authenticity models for documents, images, video and audio as generation tools improve. ## Investors URL: https://bynn.com/company/investors Bynn Intelligence has built working products and the research behind them. The next chapter is scale. Watch the film and request a private investor briefing. ### Frontier research. Working products. Our age-estimation model is #1 for true-positive rate in NIST's Child Online Safety evaluation for ages 13-16. Now we are raising capital to accelerate commercial growth. ### What Bynn has built Bynn Intelligence is a frontier AI research lab building proprietary models, decision engines and autonomous systems. That research reaches the world through a portfolio of working products and through the systems we build with enterprises. Our foundations are in fraud detection and forensic intelligence. From there, the work has grown into digital trust, age assurance, decision intelligence and autonomy. ### The portfolio Five portfolio entries under one lab, each with its own focus. - ZOREAL (Digital trust): Human trust for digital interactions: identity verification and document forensics in ZOREAL Verify, verified meetings, sign-in and signing. - Agemin (Age assurance): AI for age assurance, built on Bynn's facial age-estimation research. - Detector24 (Media forensics): Specialist models that examine what AI can create and conceal in images, video, audio and text. - DecisionNode (Decision intelligence): A decision model and inference service that returns decisions connected systems can act on. - Zora (Autonomy): A mission executive and autonomy engine for autonomous systems, demonstrated in simulation. Portfolio maturity varies. Zora is an experimental research platform shown in simulation; each page states where its entry stands. ### Evidence, labelled by type - #1 for true-positive rate in NIST's Child Online Safety evaluation for ages 13-16. True-positive rate, ages 13-16: 72.53%. Bynn's facial age-estimation model, which powers Agemin, achieved a 72.53% true-positive rate: the proportion of people aged 13 to 16 correctly estimated within that age band. Measured by NIST in its ongoing evaluation program. NIST evaluation is not certification or endorsement. The result measures age estimation within the 13 to 16 band. It does not measure a reduction in online harm. Evidence type: Independent evaluation. Source: NIST FATE Age Estimation & Verification. Submission bynn-000. https://pages.nist.gov/frvt/html/frvt_age_estimation.html (NIST results table last updated 29 September 2026). - 81.6% mean accuracy across 15 text suites in Bynn's preliminary October 2026 evaluation. Mean accuracy, 15 text suites: 81.6%. DecisionNode MoE's result in Bynn's preliminary October 2026 evaluation. The suites contain 25,134 items and cover a range of decision tasks. The published comparison includes results where other models lead. Bynn's own preliminary evaluation, run on development splits. Not an independent result. A mean across suites, not a guarantee of accuracy on any single task. Evidence type: Bynn preliminary evaluation. Source: DecisionNode benchmarks and methodology. Preliminary results, October 2026. https://decisionnode.com/benchmarks (Preliminary results, October 2026). ### Our ambition reaches further. Intelligent systems that take on difficult work and help protect people. Our mission is to advance autonomous intelligence for the benefit of humanity. - Search and rescue - Wildfire response - Energy infrastructure - Defense reconnaissance - Helping identify mines and unexploded ordnance after conflict These are application directions for our research, not deployed results. ### Reach more customers. Expand deployments. Bring more of our research into the world. Capital will accelerate commercial growth: taking the products we have built to more enterprises, and moving more of our research from the lab into use. ### Are you with us? A two-minute film about the stakes, the technology and the ambition behind Bynn Intelligence. - 0:00 The stakes - 0:12 Bynn Intelligence - 0:24 ZOREAL - 0:35 Forensic AI and Agemin - 0:50 DecisionNode - 1:00 Zora - 1:12 Application vision - 1:28 The next chapter - 1:44 Are you with us? ### Request a private investor briefing. Current traction and our plans belong in a private conversation. Request a briefing and we will walk you through the company, the portfolio and the opportunity. ## Trust Center URL: https://bynn.com/company/trust How Bynn secures enterprise AI: ISO/IEC 27001:2022 certification, a SOC 2 report, controls mapped to ISO/IEC 42001 and the EU AI Act, 72-hour breach notice. ### Trust Center Bynn Intelligence secures customer data and the models that process it, with controls monitored continuously in Vanta. #### In the contract - Breach notice within 72 hours (Data Processing Addendum §10.1) - 30 days' notice of new sub-processors (Data Processing Addendum §7.2) - Encryption in transit and at rest (Data Processing Addendum §6.1) - Transfers under the Standard Contractual Clauses (Data Processing Addendum §13) - Deletion confirmed in writing on request (Data Processing Addendum §11) ### Certifications and attestations. - ISO 27001:2022: Information security management, certified - SOC 2: SOC 2 report, monitored continuously ### Frameworks and laws. The frameworks the controls are mapped to, the laws the contract terms address and the card-data self-assessment. None of them is a certification. #### Controls mapped - ISO/IEC 42001:2023: AI management system, controls mapped to ISO/IEC 42001:2023; compliant, not certified - NIST Cybersecurity Framework: Controls mapped to the NIST CSF; a framework, not a certification - EU AI Act: Controls mapped to the EU AI Act; a law, not a certification, and no determination that any customer use complies #### Laws and self-assessment - GDPR: Processor terms under Article 28 in the Data Processing Addendum, part of every account and engagement; a law, not a certification - CCPA: Service-provider terms in the Data Processing Addendum; a law, not a certification - PCI DSS SAQ A: Card data is handled by the payment provider and never stored by us; a self-assessment, not a certification - HIPAA (BAA available): Protected health information only under a signed Business Associate Agreement, available on request; HIPAA is not a certification ### Security and privacy practices. - Encryption: Customer Personal Data is encrypted in transit over public networks and at rest. Encryption keys and credentials are managed, including their rotation. - Access control: Need-to-know access with individual accounts, and production access logged and monitored. Each customer's data is logically separated. Everyone authorised to process Customer Personal Data is bound by confidentiality and trained in handling personal data. - Sub-processors: 30 days' notice before a new sub-processor, with a right to object. Five categories: cloud computing and hosting, computing infrastructure for running models, payment processing, email delivery and customer support. Each is bound in writing to data protection obligations no less protective than the DPA, as appropriate to its services. The current list, with each sub-processor's function and location, is available to customers on request at privacy@bynn.com. - Data location and transfers: Transfers from the EEA, the UK and Switzerland under the Standard Contractual Clauses. Customer Personal Data may be processed outside the European Economic Area, the United Kingdom and Switzerland, including in the United States. The clauses are those of Implementing Decision (EU) 2021/914, with the UK Addendum for the United Kingdom and the same clauses, as adapted, for Switzerland. A request from a public authority is notified to the customer where legally permitted, challenged where unlawful and answered with the minimum information it requires. - Incident notification: Breach notice without undue delay, within 72 hours at the latest. The notice describes the breach, the people and records concerned, its likely consequences and the measures taken, with a contact point at security@bynn.com. Bynn contains, investigates and mitigates the breach, and assists with the notifications the customer must make. - Retention and deletion: Deleted when the retention period ends, and confirmed in writing on request. The retention period is the one the customer sets, or the one its order form provides. When an account closes, its data can be exported for 30 days and is then deleted. ### Responsible AI. - Decisions about people: Customers may use the Services to support decisions with legal or similarly significant effects on people only where a qualified person reviews each one before it takes effect. A person refused by an automated age or identity check must have a way to obtain that review and contest the result. - Regulated AI systems: Where a customer's use forms part of an AI system the law regulates, such as a high-risk system under the EU AI Act, the customer is its deployer or provider and carries the obligations the law places on it. Bynn provides the information about the Services that the customer reasonably requests for that purpose. - Training on customer data: Bynn uses Customer Data to develop, train or improve its models only where an order form or the product terms permit it, or, for Agemin, to improve liveness detection unless the customer opts out. Otherwise it does not. ## Compare Enterprise AI and Fraud Detection Alternatives URL: https://bynn.com/compare Alternatives to Palantir, Inscribe, Resistant AI, Shift Technology, Reality Defender and 13 more, compared with Bynn. Every vendor claim sourced and dated. ### Compare Bynn with the alternatives. Eighteen vendors in enterprise AI, document fraud, insurance claims and deepfake detection, each compared with the Bynn offering in its field. On every page: who each option suits, a side-by-side table, where they differ, common questions and the sources. Bynn is not affiliated with any vendor named. #### Enterprise AI platforms and partners Data and AI platforms, model builders and consultancies that enterprises weigh when they want AI working inside real operations. Each is compared with Bynn's enterprise AI delivery: consulting, custom development, integration, autonomous workflows and inference. - Alternatives to Palantir: Four software platforms built around the Palantir Ontology: Foundry and AIP for enterprises, Gotham for defense and intelligence, Apollo for software deployment. Private pricing only. Last reviewed 2026-10-17. https://bynn.com/compare/palantir-alternatives - Alternatives to C3 AI: An enterprise AI platform with pre-built industry applications, generative AI agents and C3 Code, started through a paid initial production deployment of generally up to six months. Last reviewed 2026-10-17. https://bynn.com/compare/c3-ai-alternatives - Alternatives to Scale AI: Training data for model builders, plus AI systems built, deployed and operated for enterprises and governments on the Scale GenAI Platform. No published price list. Last reviewed 2026-10-17. https://bynn.com/compare/scale-ai-alternatives - Alternatives to DataRobot: A platform to build, deploy and govern agents and predictive and generative models, with adoption services and forward deployed engineers. A 30-day free trial; no published price list. Last reviewed 2026-10-17. https://bynn.com/compare/datarobot-alternatives - Alternatives to Dataiku: A platform where business experts and engineers build analytics, machine learning and AI agents under central governance. A 14-day trial and a Free Edition; no published price list. Last reviewed 2026-10-17. https://bynn.com/compare/dataiku-alternatives - Alternatives to IBM watsonx: IBM's portfolio of AI products (watsonx.ai, watsonx Orchestrate, watsonx.governance) with published starting prices, alongside IBM Consulting AI services with no published rate card. Last reviewed 2026-10-17. https://bynn.com/compare/ibm-watsonx-alternatives #### Document fraud detection Specialists that check bank statements, payslips, invoices and other submitted documents for tampering, fabrication and AI generation. Each is compared with Bynn document fraud detection, whose forensics are part of ZOREAL Verify. - Alternatives to Inscribe: Agentic document fraud detection for US financial services: forensic, network, semantic and pixel-level checks on bank statements, pay stubs and tax forms, with reasoning across an application. Last reviewed 2026-10-17. https://bynn.com/compare/inscribe-alternatives - Alternatives to Resistant AI: Document fraud detection that checks PDFs and images for tampering, reuse, template farms and AI generation and returns a Trusted, Warning or High Risk verdict, plus a separate transaction monitoring product. Last reviewed 2026-10-17. https://bynn.com/compare/resistant-ai-alternatives - Alternatives to Ocrolus: A US lending document platform with extraction and cash-flow analytics, and Ocrolus Detect for document fraud, whose forensic layer Ocrolus credits to Resistant AI. Priced per document processed. Last reviewed 2026-10-17. https://bynn.com/compare/ocrolus-alternatives - Alternatives to Fortiro: Australian document fraud detection for payslips, bank statements, tax documents and invoices, sold as Protect, with income extraction (Accelerate) and redaction (Redact). Hosted on AWS in Sydney. Last reviewed 2026-10-17. https://bynn.com/compare/fortiro-alternatives - Alternatives to Klippa: Klippa, now Doxis AI.dp: a cloud document processing platform with a fraud module using EXIF, copy-move, duplicate and pixel analysis, cross-checks and an AI fraud agent. Priced on request. Last reviewed 2026-10-17. https://bynn.com/compare/klippa-alternatives #### Insurance claims fraud and automation Claims AI for insurers: fraud detection, investigations and photo-based assessment. Each is compared with Bynn's claims work: evidence collected across languages, documents and photos examined, and agentic fraud reasoning that routes the claim. - Alternatives to Shift Technology: AI agents for insurers' claims, fraud and risk work, with a cross-carrier Insurance Data Network, subrogation and payment integrity. Cloud SaaS; no published price list. Last reviewed 2026-10-17. https://bynn.com/compare/shift-technology-alternatives - Alternatives to FRISS: A Trust Automation platform for P&C insurers that scores claims, policies and customers, with Media Check for claim photos and PDFs and a case management system for SIU teams. Priced per customer. Last reviewed 2026-10-17. https://bynn.com/compare/friss-alternatives - Alternatives to Tractable: Computer vision that assesses vehicle damage from photos for insurers: total-loss triage, repair estimate pre-fill, claim review and subrogation. No published price list. Last reviewed 2026-10-17. https://bynn.com/compare/tractable-alternatives #### AI-generated media and deepfake detection Detection APIs and platforms for AI-generated images, deepfake video and cloned voices. Each is compared with Detector24, Bynn's specialist detection models for images, video, audio and text. - Alternatives to Reality Defender: A deepfake detection platform for enterprises and governments: an API, a browser review tool, and live detection for phone calls and video meetings. Free, Business and Enterprise plans published. Last reviewed 2026-10-17. https://bynn.com/compare/reality-defender-alternatives - Alternatives to Hive: AI-generated and deepfake detection APIs for images, video, audio and music, alongside Hive's moderation, generation and search models. Per-request prices published. Last reviewed 2026-10-17. https://bynn.com/compare/hive-alternatives - Alternatives to Sensity: Forensic deepfake detection for images, video and audio: face manipulation, AI-generated content, voice analysis and file forensics, with court-oriented reports and on-premise options. Last reviewed 2026-10-17. https://bynn.com/compare/sensity-alternatives - Alternatives to Sightengine: AI image, video, voice and music detection, deepfake detection and face liveness through the same API as Sightengine's content moderation. Free, Starter and Pro plans published. Last reviewed 2026-10-17. https://bynn.com/compare/sightengine-alternatives ## Research and news: Enterprise AI URL: https://bynn.com/blog Research and news from Bynn Intelligence: evaluation results, research notes and product news on document fraud, identity, age assurance, decision models and autonomy. - [Bynn ranks #1 for true-positive rate in NIST’s Child Online Safety evaluation for ages 13-16](https://bynn.com/blog/bynn-ranks-first-nist-child-online-safety): How NIST’s age estimation evaluation tests a model that guards a space for people aged 13 to 16, what the 72.53% true-positive rate counts, what it does not show, and what it means for Agemin and for enterprise buyers. - [DecisionNode MoE: 81.6% mean accuracy across 15 text suites in Bynn’s preliminary October 2026 evaluation](https://bynn.com/blog/decisionnode-preliminary-benchmark-october-2026): What Bynn’s preliminary, Bynn-run evaluation of DecisionNode measured across 15 text suites and 25,134 items, where the model leads and trails, and how an enterprise team should read an AI decision engine benchmark. - [Zora Air flies autonomous drone missions in simulation](https://bynn.com/blog/zora-air-autonomous-drone-missions-in-simulation): How Mission Control composes a mission, how Zora’s autonomy engine flies it without handing control back, what DecisionNode decides on each camera frame, and what the simulation does not yet show. - [The Future of Fraud Prevention: AI and Human Oversight](https://bynn.com/blog/the-future-of-fraud-prevention-ai-and-human-oversight): Discover how AI and human oversight help businesses detect sophisticated fraud, reduce false positives, strengthen compliance, and protect customers. - [Global Compliance Challenges & AI Verification in 220+ Countries](https://bynn.com/blog/global-compliance-challenges-ai-verification-in-220-countries): Learn how AI-powered verification helps businesses meet global KYC, AML, and fraud compliance requirements across 220+ countries and territories. - [Picture-First Document Workflows Are Creating a New Fraud Blind Spot](https://bynn.com/blog/picture-first-document-workflows-are-creating-a-new-fraud-blind-spot): Businesses often accept screenshots, scans, and phone photos because they are easier to collect. But when native documents become pictures, important integrity signals can disappear. Here is why companies now need AI that can inspect both images and PDFs in the same fraud workflow. - [How to Spot a Fake Travel Itinerary Before It Becomes a Fraud Problem](https://bynn.com/blog/how-to-spot-a-fake-travel-itinerary-before-it-becomes-a-fraud-problem): Fake travel itineraries can slip into visa files, claims, onboarding flows, and compliance reviews. Learn the red flags, verification steps, and how Bynn helps detect forged travel documents. - [AI for Insurance Fraud Detection: Reduce Risk, Cut Costs & Stop Fraud Faster](https://bynn.com/blog/ai-for-insurance-fraud-detection-reduce-risk-cut-costs-stop-fraud-faster): Discover how AI-powered fraud detection helps insurers identify fake claims, verify identities, and prevent losses. Learn how solutions like Bynn reduce risk, streamline claims, and improve compliance. - [Secure Employee Onboarding: Detect Fake Resumes with AI](https://bynn.com/blog/secure-employee-onboarding-detect-fake-resumes-with-ai): Learn how AI detects fake resumes, forged certificates, and identity fraud in hiring. Discover how Bynn helps organizations secure employee onboarding at scale. - [AI Document Forgery Detection in Under 10 Seconds](https://bynn.com/blog/ai-document-forgery-detection-in-under-10-seconds): Discover how AI detects forged IDs, passports, PDFs, and bank statements in under 10 seconds using forensic analysis, biometrics, and real-time KYC verification. - [The True Cost of Document Fraud & How to Prevent It](https://bynn.com/blog/the-true-cost-of-document-fraud-how-to-prevent-it): Discover the real cost of document fraud—from fake invoices to forged IDs—and learn how AI-powered verification from Bynn helps businesses prevent costly fraud. - [Identity Verification in Logistics: Prevent Fraud at Scale](https://bynn.com/blog/identity-verification-in-logistics-prevent-fraud-at-scale): Discover how AI-powered KYC, KYB, biometric verification, and document authentication prevent fraud in transport and logistics without slowing operations. - [KYC/AML Best Practices for Real Estate in 2026](https://bynn.com/blog/kyc-aml-best-practices-for-real-estate-in-2026): Discover KYC/AML best practices for real estate firms in 2026. Learn how identity verification, KYB, and AML monitoring reduce fraud and regulatory risk. - [KYC Onboarding Platforms for Startups](https://bynn.com/blog/kyc-onboarding-platforms-for-startups): Learn how startups scale KYC onboarding without slowing growth—AI document checks, biometrics, liveness, fraud signals, and automated risk decisions with Bynn. - [AI-Generated Fake Documents in 2026: Fraud Risks & Prevention](https://bynn.com/blog/ai-generated-fake-documents-in-2026-fraud-risks-prevention): AI-generated fake documents are reshaping fraud in 2026. Learn why traditional checks fail and how businesses can stay ahead with modern detection strategies. - [Real-Time Online ID Verification Services for 2026](https://bynn.com/blog/real-time-online-id-verification-services-for-2026): Discover how Bynn’s real-time ID verification speeds up onboarding, cuts fraud, and ensures KYC/AML compliance across industries in 2026. - [Seamless KYC and Customer Onboarding](https://bynn.com/blog/seamless-kyc-and-customer-onboarding): Accelerate customer onboarding with Bynn’s AI-powered KYC tools. Reduce fraud, stay compliant, and build trust from the first click. - [Top Identity Verification Measures to Prevent Fraud in 2026](https://bynn.com/blog/top-identity-verification-measures-to-prevent-fraud-in-2026): Explore the top identity verification tools businesses need in 2026 to fight fraud, meet KYC/AML rules, and onboard users securely with Bynn’s AI solutions. - [Structuring in Money Laundering: Risks and How Bynn Detects It](https://bynn.com/blog/structuring-in-money-laundering-risks-and-how-bynn-detects-it): Learn how structuring enables money laundering and how Bynn’s AI tools detect and stop it. Stay compliant and secure with real-time fraud prevention. - [Customer Risk Assessment for Compliance & Fraud Prevention](https://bynn.com/blog/customer-risk-assessment-for-compliance-fraud-prevention): Discover how customer risk assessment enhances KYC, KYB, and AML compliance, prevents fraud, and learn how AI-driven platforms streamlines the process. - [What Is a Document Verification Software? Benefits, Use Cases & Features](https://bynn.com/blog/what-is-a-document-verification-software): Discover what a document verification software is and how Bynn uses AI & biometrics to speed up onboarding, prevent fraud, and ensure KYC/KYB compliance. - [AML and Fraud Detection Solutions for Safer Business Operations](https://bynn.com/blog/aml-and-fraud-detection-solutions-for-safer-business-operations): Discover how Bynn’s AI-driven AML and fraud detection tools protect your business from financial crime, compliance risk, and identity fraud. - [Inside the Mind of a Fraudster: Money Laundering Schemes and Examples](https://bynn.com/blog/inside-the-mind-of-a-fraudster-money-laundering-schemes-and-examples): Explore real-world money laundering schemes and learn how Bynn’s AI-driven KYC and AML tools help businesses detect and prevent financial crime. - [What is Enhanced Due Diligence (EDD)? A Guide to EDD](https://bynn.com/blog/what-is-enhanced-due-diligence-edd-a-guide-to-edd): Discover what Enhanced Due Diligence (EDD) is, why it matters for compliance, and how Bynn helps businesses streamline EDD processes and reduce risks. - [Sanctions Screening Explained: Stay Compliant](https://bynn.com/blog/sanctions-screening-explained-stay-compliant): Learn how PEP screening helps identify politically exposed persons and meet AML compliance using Bynn’s AI-powered identity verification tools. - [AML Risk Assessment: Spot and Reduce Financial Crime Risks](https://bynn.com/blog/aml-risk-assessment-spot-and-reduce-financial-crime-risks): Learn how to perform effective AML risk assessments and how Bynn’s AI tools help reduce exposure to money laundering and regulatory penalties. - [How to Find Out if a Company is Legit: Tips for Safe KYB](https://bynn.com/blog/how-to-find-out-if-a-company-is-legit-tips-for-safe-kyb): Discover essential tips to verify a company’s legitimacy. Protect yourself from scams and make informed decisions. Read the article for practical advice. - [Why Identity Verification for Online Dating is important in 2025](https://bynn.com/blog/why-identity-verification-for-online-dating-is-important-in-2025): Prevent catfishing and romance scams with Bynn’s AI-powered identity verification. Build safer, more trusted dating platforms in 2025. - [Global Overview of Mandatory QR Codes on Invoices and Receipts](https://bynn.com/blog/global-overview-of-mandatory-qr-codes-on-invoices-and-receipts): In today’s digitized economy, a small square QR code on an invoice or receipt can carry huge significance. Around the world, tax authorities are mandating the use of QR codes in invoices and receipts to combat fraud, streamline audits, and improve tax compliance. For multinational enterprises, the rapid spread of QR in invoice and receipts requirements means navigating new regulations in nearly every market – and the stakes are high. Embracing this trend is not just about avoiding fines; it’s about leveraging technology to enhance transparency and efficiency in financial operations. - [What Is PEP Screening? Identifying Politically Exposed Persons for AML Compliance](https://bynn.com/blog/what-is-pep-screening-identifying-politically-exposed-persons-for-aml-compliance): Learn how PEP screening helps identify politically exposed persons and meet AML compliance using Bynn’s AI-powered identity verification tools. - [Anti-Money Laundering Examples: Prevent & Protect Your Business](https://bynn.com/blog/anti-money-laundering-examples-prevent-protect-your-business): Explore real-world anti-money laundering scenarios and learn how proactive AML strategies and advanced solutions like Bynn can safeguard your business effectively. - [How To Prevent Fraud: Best Practices to Protect Your Business](https://bynn.com/blog/how-to-prevent-fraud-best-practices-to-protect-your-business): Discover practical fraud prevention strategies, from advanced KYC/KYB checks to AI-powered detection, to protect your business effectively with Bynn. - [Economic Crime and Corporate Transparency Act (ECCTA) Solution: Ensuring Compliance with AI-Powered Verification](https://bynn.com/blog/economic-crime-and-corporate-transparency-act-eccta-solution-ensuring-compliance-with-ai-powered-verification): The Economic Crime and Corporate Transparency Act (ECCTA) is a landmark UK law enacted in 2023 to enhance corporate transparency and combat economic crime. It strengthens regulations to ensure that companies are not used for illicit purposes, chiefly by introducing stricter identity verification and reporting requirements. In essence, ECCTA mandates that key individuals associated with UK companies prove they are who they claim to be. This includes a new mandatory identity verification system for people involved in running or controlling companies. Specifically, all company directors and persons with significant control (PSCs) – both new and existing – are now required to verify their identities with the companies registry. Even individuals who file documents on a company’s behalf must be verified, unless they’re registered agents under the Act. - [The Real ID Act in 2025: Compliance, Enforcement, and What It Means for Identity Verification](https://bynn.com/blog/real-id-act-for-identity-verification): The Real ID Act is a U.S. federal law passed in 2005 as a post-9/11 security measure. It was enacted in response to recommendations by the 9/11 Commission to establish federal standards for identification documents. In essence, the Act sets nationwide security requirements for state-issued IDs (like driver’s licenses and identification cards) to make them more secure and harder to forge. This law aims to enhance national security by ensuring that state IDs can be reliably used for federal purposes (such as passing through airport security) without fraud or identity tampering. - [Stop Synthetic Identity Fraud with Real-Time Government Verification](https://bynn.com/blog/stop-synthetic-identity-fraud-with-real-time-government-verification): Synthetic identity fraud is an invisible threat infiltrating businesses at alarming rates. Fraudsters create fake personas by stitching together real and fictitious information – a “Frankenstein” identity – to open accounts and commit crimes undetected. The result? The fastest-growing form of identity theft, costing businesses billions globally. In the U.S. alone, the average loss per synthetic fraud case is around $15,000, and shockingly over 80% of new account fraud today is attributed to synthetic identities. This isn’t a niche issue; it’s a growing epidemic. Nearly half of organizations worldwide experienced synthetic identity fraud in the past year, and 91% of U.S. companies see these fake identities as a growing threat. Looking at recent trends, suspected synthetic fraud incidents rose 21% in just one year (184% since 2019). In short, synthetic identities have become a favorite weapon of cybercriminals – and no industry is immune. - [Navigating Age Verification Laws in the USA and Europe: Compliance for iGaming, Social Media, and Beyond](https://bynn.com/blog/age-verification-compliance-igaming-social-media): Age verification has become a critical compliance checkpoint for online businesses – from iGaming platforms and online casinos to social media networks and content sites. Lawmakers in the United States and Europe are increasingly focused on keeping minors away from age-restricted content and services. This means companies must verify users’ ages and meet strict legal standards, all while respecting user privacy. In this article, we’ll explore the key age verification laws in the U.S. and Europe (including regulations like GDPR, COPPA, KYC requirements, and the EU Digital Services Act) and what they mean for industries like online gambling and social media. We’ll also discuss whether businesses are required to use specialized age verification tools (such as biometric age-check SDKs like Bynn.com), and how privacy-preserving methods – like facial age estimation and zero-knowledge proofs – can confirm a user’s age without storing sensitive data. Finally, we’ll clarify the difference between age verification and full identity verification, and why each serves a different purpose. - [Optimizing KYC Verification: Balancing FRR and FAR Across Industries and Regions](https://bynn.com/blog/optimizing-kyc-verification-balancing-frr-and-far-across-industries-and-regions): Ever had a legitimate customer blocked by your KYC system even though they were who they claimed? That’s an example of a false rejection. Or worse, imagine a fraudster slipping through as a verified user – a false acceptance. In identity verification, these two error metrics are critical. The False Rejection Rate (FRR) measures how often a good user is mistakenly rejected, while the False Acceptance Rate (FAR) measures how often a bad actor is wrongly accepted. - [Guide to KYC Compliance for Crypto Exchanges and Wallets](https://bynn.com/blog/guide-to-kyc-compliance-for-crypto-exchanges-and-wallets): Know Your Customer (KYC) is the process of verifying customer identities and background to meet Anti-Money Laundering (AML) requirements. In the crypto industry, KYC has become essential for exchanges and wallet providers to operate legally and safely. KYC for crypto typically involves a set of steps during user onboarding to confirm who the customer is and assess their risk – steps which are legally required under financial regulations. By implementing robust KYC procedures, crypto platforms can deter illicit activity; vigorous identity checks make it much harder for criminals to launder money or commit fraud through an exchange, effectively dissuading bad actors from using that service. This not only protects the business and its users from fraud and theft, but also builds trust, allowing crypto exchanges and wallets to integrate with the broader financial system. In short, strong KYC compliance helps legitimize the crypto industry, making it safer and more credible for everyday users and regulators alike. - [Digital Identity Verification in Banking vs Fintech: A Global Perspective](https://bynn.com/blog/digital-identity-verification-in-banking-vs-fintech-a-global-perspective): Financial institutions worldwide face a critical balancing act in digital identity verification (IDV). Traditional banks and fintech startups both must verify customer identities to meet Know Your Customer (KYC) and Anti-Money Laundering (AML) regulations, but they often approach the task differently. This thought leadership overview compares how legacy banks and agile fintechs tackle IDV – from compliance hurdles and onboarding workflows to the latest innovations in AI-driven verification. We also explore global regulatory landscapes and fraud prevention strategies, highlighting key differences in approach. - [KYC in Crypto](https://bynn.com/blog/kyc-crypto): Cryptocurrencies have revolutionized finance with promises of decentralization and anonymity. However, as the industry matures, it faces growing scrutiny and regulation. Know Your Customer (KYC) procedures have become a cornerstone of compliant and trustworthy crypto services. In simple terms, KYC means verifying the identity of users. This article breaks down what KYC is and why it’s essential for crypto, how KYC works on exchanges and other blockchain platforms, the regulatory requirements behind it, benefits for both individuals and businesses, common challenges, and why having a reliable identity verification solution (like bynn.com) is so important for businesses in the crypto space. - [Onboard Customers Smoothly: Reducing Users’ Drop-off Rate in Identity Verification](https://bynn.com/blog/onboard-customers-smoothly-reducing-users-drop-off-rate-in-identity-verification): Identity verification is a critical checkpoint in user onboarding – but it’s also a common point where users abandon the process. Losing users at this stage (“drop-off”) can be costly, wasting the marketing and acquisition efforts that brought them to sign up in the first place. This article examines why users drop off during KYC (Know Your Customer) or KYB (Know Your Business) checks and outlines best practices to keep them engaged. We’ll also look at how Bynn.com’s identity verification services apply these best practices to improve completion rates, and review some data on how streamlining verification boosts user retention. - [The Future of Identity Verification: Trends & Challenges in 2025](https://bynn.com/blog/the-future-of-identity-verification-trends-challenges-in-2025): Digital identity verification has become a cornerstone of the modern digital economy, underpinning everything from banking to government services. As we enter 2025, the landscape of remote identity verification is undergoing a major transformation, driven by rapid technological advancements, increasing fraud threats, and tightening regulations. This evolution aims to meet growing security and compliance needs while maintaining a seamless user experience. The following report explores key trends shaping digital identity verification, the challenges that lie ahead, industry-specific applications, and a forward-looking outlook with recommendations for businesses and regulators. - [Understanding KYB (Know Your Business) – A Guide to Business Verification](https://bynn.com/blog/understanding-kyb-know-your-business-a-guide-to-business-verification): Understanding KYB (Know Your Business) – A Guide to Business VerificationIntroductionIn today’s global business environment, compliance and trust are more crucial than ever. Financial crimes like money laundering and fraud are on the rise – an estimated $800 billion to $2 trillion is laundered globally each year – pushing regulators to tighten oversight. This is where Know Your Business ( - [KYC vs AML vs KYB: Key Differences](https://bynn.com/blog/kyc-vs-aml-vs-kyb-key-differences): Money laundering and related financial crimes pose a serious global threat, with an estimated $800 billion to $2 trillion laundered worldwide each year. To combat this, regulators worldwide enforce strict compliance frameworks: Know Your Customer (KYC), Know Your Business (KYB), and Anti-Money Laundering (AML). These frameworks work together to verify customer identities, vet business entities, and detect illicit financial activity, thereby protecting the integrity of the financial system. Businesses are legally required to comply with KYC, KYB, and AML standards – failing to do so can result in heavy fines and reputational damage. This guide explains each concept, highlights their differences and overlaps, and outlines best practices for global compliance (with a focus on how services like bynn.com can assist in meeting these obligations). - [Risk-Based Approach Best Practices Guide 2025](https://bynn.com/blog/risk-based-approach-best-practices-guide-2025): A Risk-Based Approach (RBA) in compliance means understanding the specific risks your organization faces and tailoring controls proportionately to those risks. Instead of a one-size-fits-all checklist, an RBA allocates more resources to higher-risk areas and fewer to low-risk areas, making compliance efforts more efficient and effective. In financial services and fintech, this approach is crucial – global regulators like the Financial Action Task Force (FATF) have made RBA central to anti-money laundering (AML) recommendations, reinforcing that effective compliance must prioritize the most severe threats. - [What is Identity Verification? How It Works & Why It’s Crucial in 2025](https://bynn.com/blog/identity-verification-idv-in-2025-a-comprehensive-beginner-friendly-guide): Identity Verification (IDV) has become a cornerstone of doing business in the digital age. Verifying that someone is who they claim to be is crucial for preventing fraud, complying with laws, and building trust online. This guide provides a beginner-friendly overview of IDV in 2025 – why it matters, how it works, where it’s used, and what the future holds – tailored for businesses, iGaming operators, and financial institutions around the world. - [AI Makes Creating Fake IDs Easier Than Ever](https://bynn.com/blog/ai-makes-creating-fake-ids-easier-than-ever): Advances in artificial intelligence—especially generative adversarial networks (GANs) and deepfake technology—have dramatically lowered the barrier to producing convincing fake identification documents. Generative AI models can now fabricate realistic driver’s licenses, passports, and other IDs with photorealistic faces, seals, and text, often indistinguishable from genuine documents at first glance.