Arkose Labs Expands Agent Controls for Businesses
The expanded Arkose Agent Trust Manager lets businesses classify and set policies for AI agent sessions as agents increasingly fail to identify themselves.
Updated on Nov. 5, 2026 in Artificial Intelligence
Arkose Labs has expanded its Arkose Agent Trust Manager to govern agent sessions, letting businesses classify each session and apply allow, monitor, or block policies at every step of the customer journey. The move follows findings from its State of Agentic AI Trust Q3 report that personal AI agents identify themselves less than 3% of the time.
Why it matters
Businesses can no longer count on AI agents to announce who they are, making it harder to distinguish legitimate automated activity from malicious traffic. Tools that classify and track agents across networks aim to restore visibility into automated interactions.
The State of Agentic AI Trust Q3 report found personal AI agents identify themselves less than 3% of the time, a figure drawn from Arkose Labs' own analysis. Sampling methodology and comparison periods were not provided in the source extraction.
The players
Arkose Labs
Arkose Labs is a company that provides bot detection and fraud prevention tools for online businesses.
The details
Arkose Agent Trust Manager classifies every agent session as self-disclosing good, non-disclosing good, or malicious. Arkose Labs is also introducing Agent ID, a stable identifier that follows each agent rather than its IP address, along with detection auditing exportable from Command Center. Blocked agents receive a brand-specific message stating the workflow is not supported for agents.
Timeline
Q3: Arkose Labs publishes the State of Agentic AI Trust Q3 report finding agents self-identify under 3% of the time.
Coming soon: A proof-of-human feature launches requiring a person to confirm before an agent can continue on sensitive steps.
Industry Landscape
Arkose Labs' Agent Trust Manager extends its existing bot-detection products into classifying AI agents by session and step of the customer journey. The expansion responds to a shift in which automated agents no longer reliably self-identify to the sites they access.
Businesses can set allow, monitor, or block rules for AI agents at checkout, login, and sign-up, which can affect how automated tools interact with online services. Users whose agent workflows are blocked will see a notice that the workflow is not supported for agents.
The takeaway
As AI agents act on users' behalf online, businesses are shifting from trusting self-identification to classifying and tracking sessions themselves. Reviewing how a service handles automated access can clarify what tools and workflows may be affected.
What happens next
Arkose Labs is launching a proof-of-human feature that will require a person to confirm before an agent can continue on sensitive steps, with no date specified.
Further reading
Explore more on Artificial Intelligence developments and the tools shaping automated online activity.
