Incognia launches AI Agent Detection to secure autonomous financial transactions. Incognia has launched AI Agent Detection to help financial institutions verify whether AI-initiated transactions are actually authorized by the human account holder. As agentic AI begins managing payments and account changes, this capability addresses the critical gap between identifying a legitimate bot and confirming the underlying user's intent, protecting against unauthorized automated activity. What was announced. Incognia has introduced AI Agent Detection, a tool designed to help financial institutions differentiate between legitimate AI-driven actions and unauthorized automated requests. The system addresses a growing vulnerability where AI agents, acting on behalf of consumers, perform sensitive tasks like making purchases or changing account settings. The capability works by evaluating the agent and the specific action as two distinct entities. It utilizes signed-request verification, agent-origin classification, and bot detection, which are then cross-referenced with Incognia's broader risk intelligence. To increase confidence in high-risk transactions, the platform can verify actions against signals from a customer's trusted mobile devices, historical account activity, and network location. Alongside this, Incognia expanded its web risk intelligence through Web Behavioral Biometrics. This adds new signals to its browser and device intelligence, including mouse movements, keyboard interactions, and copy-and-paste activity. If a web session appears suspicious, the system can link it to a trusted mobile device to confirm the user's identity. Furthermore, the company launched an MCP Server to integrate this intelligence into AI-assisted fraud investigations. This allows fraud analysts to use compatible AI applications to retrieve evidence and identify connections across accounts using natural-language queries. To ensure security, the initial MCP experience is read-only, maintaining human control over final enforcement decisions. These tools are aimed at global companies across mobility, marketplaces, and financial services. "AI agents change the trust problem for financial institutions. Knowing that a request came from a legitimate agent isn't the same as knowing that a specific action should be allowed. Businesses need to connect that request back to the real customer and the context surrounding the action so they can support legitimate agentic experiences without giving up the risk controls they've spent years building." André Ferraz, CEO and Co-Founder at Incognia. The companies involved. Incognia is a risk intelligence provider specializing in cross-device authentication and fraud prevention. The company distinguishes itself in the market through the use of apartment-level location intelligence, which it combines with device and behavioral signals to recognize trusted users across different sessions and devices. Its technology is designed to support global enterprises across several sectors, including financial services, mobility, and digital marketplaces. By focusing on high-precision location and device integrity, the firm aims to reduce the friction legitimate users face while blocking sophisticated fraudulent activity. The company's approach centers on the concept of "trusted environments," moving away from static passwords or easily intercepted one-time codes. In the context of the current announcement, Incognia is positioning itself as a governance layer for the emerging "agentic economy," where software agents act as intermediaries for human users. Supporting this industry shift is Datos Insights, a firm represented by Jim Mortenson, who serves as a Fraud Strategic Advisor. Datos Insights provides analysis and advisory services to the financial industry, focusing on the evolution of control infrastructures and the necessity of distinguishing between human and automated actors. What FF News has reported before. FF News has previously tracked Incognia's efforts to map the evolving landscape of automated and industrial-scale fraud. In the report Incognia Exposes How Overseas Scam Cities Target Global Financial System, published in September 2026, the company detailed the rise of "fraud farms." That investigation highlighted how organized criminal entities utilize specialized infrastructure to target the global financial system, often operating out of specific geographic hubs to bypass traditional security perimeters. This prior research underscores the company's focus on identifying the origin and integrity of digital interactions, a theme that continues with its latest focus on agentic AI detection and web behavioral biometrics. What this means. The introduction of AI-specific detection marks a pivot in the cybersecurity industry from "identity verification" to "intent verification." As consumers delegate financial authority to AI agents, the traditional security perimeter - which relies on a human being physically interacting with a screen - is becoming obsolete. This announcement puts pressure on traditional fraud vendors who rely on legacy multi-factor authentication (MFA) that agents cannot easily navigate. The industry now faces a critical question: how to maintain a "human-in-the-loop" for high-value transactions without destroying the efficiency that AI agents are supposed to provide. This move suggests that the future of fintech security will depend less on what a user knows and more on the verifiable origin of the automated request. Featured speakers.

Incognia launches agent checks that tie payments to customer context. Co-founder André Ferraz is extending Incognia's device, location and behavioral fraud signals to agent-driven actions, web sessions and analyst investigations. By RuntimeWire Staff - Published Oct 6th, 2026, 12:20pm CT - 4 MIN READ - 5 SOURCES - VIA PR Newswire Listen to article. Incognia introduced a system for financial institutions to assess whether an AI agent is legitimate and whether the action it is attempting is authorized by the customer behind it. The October 6th launch builds on co-founder and CEO André Ferraz's use of device, location and behavioral signals to establish trust in digital interactions. The new AI Agent Detection capability evaluates the agent and its requested action separately. Incognia says it combines signed-request verification and agent-origin classification with bot detection and its existing risk intelligence. When a bank needs more confidence, it can compare an agent's request with signals from the customer's trusted devices, account activity, device integrity, network and location. That distinction addresses a specific problem in agent-driven payments: confirming that a request came from a real agent does not establish that its human customer authorized a particular purchase, payment or account change. Ferraz put it plainly in the launch announcement: "Knowing that a request came from a legitimate agent isn't the same as knowing that a specific action should be allowed." How Incognia applies location technology to agentic commerce. Ferraz and co-founder and CTO Alan Gomes began developing the location technology behind Incognia around 2010. The company's account of its history describes early work on indoor location before the founders turned that foundation toward digital identity and fraud prevention. Ferraz studied computer science at the Federal University of Pernambuco; Gomes studied there as well, and both were co-founders of the predecessor company In Loco, according to Incognia's team history. Incognia aims to identify whether a digital interaction fits a user's established device and location patterns, including when a person changes devices or sessions. For agentic commerce, Ferraz is applying that approach to a different question: can the institution connect an automated request to the person whose authority it invokes? The launch also bundles two adjacent capabilities. Web Behavioral Biometrics adds mouse movement, keyboard interaction and copy-and-paste behavior to Incognia's browser, device, automation and location signals. The company says it can connect a web session to a trusted mobile device when further verification is needed. Its MCP Server, meanwhile, is designed to let compatible AI applications retrieve fraud evidence, trace links across devices and accounts, and organize investigation findings from natural-language questions. The initial MCP experience is read-only, leaving enforcement decisions with human analysts. The read-only MCP experience keeps AI in the role of gathering and organizing evidence. A person retains decisions such as changing a risk outcome or taking enforcement action. In a regulated fraud operation, the audit trail and the decision-maker are part of the product. Incognia joins other efforts to establish trust in agentic commerce. Incognia is entering a field where identity vendors and payment providers are building ways to connect people with software agents. Experian announced its Agent Trust framework in April 2026, describing a system to establish a verifiable link between consumers and agents for AI-driven commerce. IBM announced an MCP-based capability for fraud detection and investigation in March. Those launches address overlapping trust problems, although their products and approaches are not directly comparable based on the public descriptions. Incognia says its products focus on the risk around a specific consumer action and the context available across that customer's devices and behavior. This builds on its existing fraud-prevention business, which combines location analysis, device recognition and tamper detection. The new products extend those signals into agent requests, browser sessions and investigators' AI workflows. The launch announcement does not provide customer deployments or adoption figures for these capabilities. Incognia announced a $31 million Series B in January 2024, led by Bessemer Venture Partners, with participation from FJ Labs and existing investors including Point72 Ventures, Prosus and Valor Capital, according to its funding announcement. That round predates this launch by more than two years and is financing history, not new capital for the products announced today. Banks will need to determine whether signals they already understand - known devices, account history, location and behavior - can help distinguish authorized automation from a legitimate agent acting outside its customer's intent. Incognia has described the controls it wants to offer, but has not supplied deployment figures or independent performance results for the new capabilities. Their operational value remains to be established in use. RuntimeWire, Inc. get virtually zero search traffic. Please consider sharing this story. Signal check Noise Signal Where does this story land? Drag the dial. The startup Incognia helps companies stop fraud using location, persistent device recognition, and tamper detection. Incognia provides fraud detection and identity intelligence for platform businesses. Its core product, Incognia ID, combines app security, device identity, and location intelligence to recognize users across devices and detect organized fraud involving resets, clones, and spoofed environments. The company serves food delivery, ride sharing, P2P marketplaces, and financial services, addressing problems such as account takeover, promo and refund abuse, fake accounts, payment fraud, and account sharing. Its site says more than 50% of top delivery platforms trust Incognia and reports a 0.001% false positive rate. Named customers and partners include Zepz, Webull, Nextdoor, Delivery Hero, Just Eat Takeaway, Grubhub, Favor, Cabify, and SumUp. Stay on the wire. Breaking startup & AI news, funding scoops, and launches as they happen. Follow @runtimewire on Threads and LinkedIn.