The platform evaluates agents and their actions as distinct entities. By combining signed-request verification with device integrity, network signals, and historical account behavior, the system validates whether a specific AI command aligns with the customer's established patterns. According to André Ferraz, Incognia’s CEO, businesses must bridge the gap between recognizing a legitimate agent and verifying that a specific transaction is actually desired by the human user.
Incognia Debuts AI Agent Detection to Secure Financial Transactions
Financial institutions are struggling to distinguish between authorized consumer activity and malicious AI-driven requests. San Jose-based Incognia launched a new detection suite today designed to verify the intent behind agentic actions, ensuring that legitimate AI convenience does not bypass critical security controls or facilitate unauthorized account changes.
Beyond agent oversight, Incognia expanded its web intelligence with behavioral biometrics, tracking mouse movements, keystrokes, and copy-paste activity to identify potential spoofing. The company also introduced an MCP Server, allowing fraud analysts to query risk data using natural language within AI-assisted investigation tools. This read-only interface ensures that while AI helps organize evidence, the final enforcement decisions remain firmly under human control.
Comments (0)
No comments yet. Be the first!