Socure Launches Agent Intelligence to Score Identity and Trust for AI Agents

Identity and risk intelligence company Socure released Agent Intelligence on October 27, 2026, designed to identify AI agents, score their trustworthiness, and link each agent to a verified individual or business. According to the company's announcement, the tool achieved 99.99% accuracy in controlled lab tests when identifying named consumer-grade agents such as Muse and Instinct.

Socure says most institutions currently cannot distinguish traffic coming from new consumer-grade agents: they cannot see them, do not know how many there are, and are unclear what they are doing. Most existing verification methods at scale were designed to verify real people, while each agent adds another layer between the behavior and the person behind it.

The product's scoring system is divided into three components—identity, authorization, and risk—with each score returned separately so customers can prioritize metrics according to their own use cases. Founder and CEO Johnny Ayers said that blocking everything would turn away the next customer, while allowing everything is equivalent to handing the keys to fraudsters; the way forward lies in determining whether something is an agent, which agent it is, and who is behind the agent.

A previous PYMNTS Intelligence report showed that nearly 90% of organizations list bot management as a challenge, and 52% said bot traffic increased in the 12 months through September 2025.

Source: https://www.pymnts.com/news/artificial-intelligence/2026/socure-launches-tools-to-determine-ai-agents-trustworthiness/