Photo: Steve A Johnson / Pexels
The new capability helps government agencies spot when AI agents, not people, are logging into public services
Government websites were built for people. Increasingly, the visitors are software acting on their behalf.
LexisNexis Risk Solutions has announced new capabilities called Know Your Agent (KYA) intelligence. The tool helps government agencies identify when an artificial intelligence agent enters or interacts with a digital government system. The company says it treats that agentic activity as a distinct risk signal, one that can be weighed alongside everything else an agency already watches.
The timing matters. Agentic traffic grew by over 750% in the first half of 2026. Systems designed around the assumption of a human at the keyboard now have to account for a very different kind of user.
What Know Your Agent actually does #
KYA is aimed at the parts of government where mistakes get expensive. LexisNexis lists citizen accounts, public benefits, tax systems and licensing portals among the digital services it is designed to protect.
The capability is not a standalone product. It is available through two existing LexisNexis platforms: ThreatMetrix and BehavioSec.
That integration is the main selling point. Agencies can fold AI-agent detection into the same risk workflows they already use for digital identity, device, behavioral and transaction intelligence. Instead of bolting on a separate AI alarm, they get one more input feeding a familiar dashboard.
Under the hood, the system looks for AI-agent signatures using signals such as browser and device data. The goal is to separate three groups that can look similar from the outside: humans, traditional bots and AI agents.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
The scale behind the signals #
Detection tools are only as good as the data that trains and informs them. LexisNexis says its Digital Identity Network analyzed over 116 billion digital transactions in 2025.
The government launch, announced on November 10, 2026, follows a commercial version announced on November 5, 2026. According to the research findings, many organizations are already incorporating KYA signals into their live risk workflows.
Mike Woodberry of LexisNexis Risk Solutions framed detection as only the first step.
“Managing risk in the agentic economy will require more than agent detection alone. The real challenge is establishing context…”
Why agencies cannot simply block the bots #
The company says KYA is meant to protect public services while still enabling legitimate uses of AI. Some citizens will reasonably want an assistant to renew a license or check a benefits application. At the same time, the same automation that helps an honest user can help a fraudster scale up. An agent that fills out one form on behalf of a person could, in the wrong hands, fill out thousands. Public benefits and tax systems are exactly the kind of targets where that difference carries real consequences.
The product’s focus on policy-based decisions for sensitive services is the middle path. Agencies can set rules based on context. An agent checking a refund status might be fine. An agent changing the bank account attached to that refund might warrant extra scrutiny, or a human in the loop.
What this means for agencies and the identity industry #
The name is doing some work. “Know Your Agent” deliberately echoes the identity verification checks that financial firms run on customers. The message is that AI agents may soon need their own version of the identity checks long applied to people.
For government agencies, the practical shift is conceptual. Risk teams have long sorted traffic into “probably human” and “probably malicious bot.” An agent acting legitimately for a verified citizen fits neither bucket cleanly, and existing rules may either block it unfairly or wave it through without question. The commercial release came first, and the government version followed within days.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.