Sumsub has launched a new AI Agent Verification solution designed to establish clear human accountability behind AI-driven automation, addressing growing concerns around fraud, anonymity and trust in digital systems.
Unveiled in Singapore, the solution introduces what Sumsub describes as a first-of-its-kind “agent-to-human binding” framework, linking AI agents directly to verified human identities within the company’s Know Your Agent (KYA) model. The approach aims to allow legitimate automation to operate at scale while preventing malicious activity driven by anonymous or unaccountable AI agents.
As AI agents and browser-based automation become more widely used, many platforms struggle to distinguish legitimate automated activity from fraud. In response, organisations often block automation entirely, creating friction for lawful users and businesses. Sumsub said its AI Agent Verification framework enables a more nuanced, risk-based approach by tying every automated action to a real, verified individual.
Rather than treating automation as an automatic threat, the system first detects whether activity is automated and then assesses its risk level. Additional verification is only applied when risk thresholds are exceeded. In higher-risk scenarios, the platform can require a targeted liveness check to confirm that a real human is present and authorised, preventing deepfakes or impersonation from being used in place of legitimate users.
Sumsub co-founder and chief technology officer Vyacheslav Zholudev said AI agents are rapidly becoming central to digital operations but are often treated as opaque and unaccountable. He said the company’s approach focuses on verifying the humans behind AI agents, rather than attempting to place trust in the agents themselves.
The AI Agent Verification capability builds on Sumsub’s broader full-cycle verification platform, combining device intelligence and bot detection to identify automated activity in real time, mule network prevention to detect coordinated fraud across devices and accounts, and continuous risk scoring throughout the customer lifecycle. Liveness verification can be triggered at critical moments such as onboarding, account changes or high-value transactions to ensure accountability.
According to Sumsub’s Identity Fraud Report 2025–2026, AI-driven fraud has grown significantly, with a 180 per cent year-on-year increase in multi-step, coordinated attacks globally in 2025. The report highlights the increasing use of AI agents in sectors such as fintech, payments, e-commerce and ticketing, where automation is becoming unavoidable as organisations pursue efficiency at scale.
Sumsub head of fraud prevention Artem Popov said the core issue is no longer automation itself, but anonymity. He said that when AI agents can move funds or transact at scale without a real person being accountable, fraud becomes far harder to control. By requiring verified human responsibility at high-risk points, he said AI Agent Verification changes the dynamic and restores trust.
Sumsub said the new capability is designed to help organisations enable trusted automation while maintaining strong safeguards against AI-driven fraud, ensuring that every automated action can be traced back to a verified individual.
