Sumsub expands Know Your Agent framework to control AI access

Sumsub has expanded its Know Your Agent (KYA) framework with the launch of the KYA Module for Model Context Protocol (MCP). This update allows businesses to define specific access permissions for AI agents, whether they are developed in-house or built using models such as Claude or ChatGPT.

The module acts as a filter between an AI agent and connected data, screening requests and stripping personally identifiable or sensitive information before it reaches the agent. It utilizes a configurable rule engine that enables precise control based on data sensitivity, the type of action, and the requesting or calling departments.

This expansion follows previous milestones, including the January launch of AI Agent Verification, which links automation to verified human identities, and the June integration of MCP. The move addresses security concerns regarding autonomous agents; Sumsub notes that 65% of organizations experienced at least one security incident caused by an agent on their network in the past year, often involving sensitive data exposure.