Arkose Agent Trust Manager Now Lets Businesses Orchestrate What Every AI Agent is Allowed to Do, Step by Step

New Identification, Classification and Policy Capabilities Enable Companies to Allow, Monitor, or Block Agents at Each Step, Driving Revenue and Preventing Fraud

SAN MATEO, CA – Nov. 5, 2026 – Arkose Labs, the leader in agent trust and orchestration, today announced a major expansion of Arkose Agent Trust Manager that shifts businesses from simply seeing agentic traffic to controlling it. Companies can now: classify every agent session; set allow, monitor, or block policies at each step of the customer journey; and follow the same agent across multiple networks. The need is urgent given Arkose Labs’ State of Agentic AI Trust Report, Q3 2026 found that personal AI agents identify themselves less than 3% of the time. Rules that wait for an agent to announce itself miss almost all of them. These controls can be managed, combined with additional intelligence sources, and shared with internal teams via the Command Center.

“Agents are already changing how they identify themselves to avoid being blocked, and they can pass authentication steps built for people. Businesses can no longer count on agents to announce who they are,” said Kevin Gosschalk, founder and CEO of Arkose Labs. “With Agent Trust Manager, they decide what every agent is allowed to do at each step of the customer journey, and those rules hold even when an agent changes its signals or its network.”

New Agent Trust Manager capabilities include:

• Agent Orchestration: Businesses lay out their customer journey step by step, sign-up, login, checkout, money movement, and set a policy for each: allow, monitor, or block. Arkose Titan applies it while the interaction happens.
• Agent Visibility and Classification: Every agent session is classified as self-disclosing good, non-disclosing good, or malicious, so businesses know which agents are in their flows and whether each one is acting for a real customer.
• Agent Rules and Agent Block Message: Enforce the policy and tell the agent why: a blocked agent sees a clear message in the business’s brand that the workflow is not supported for agents, with a pointer to the path that is.
• Ask Titan and AI Assist Integration (MCP): Plain-language answers about the agents in a business’s traffic, in Command Center or in the AI assistant it already uses, governed by existing roles.
• Introducing Arkose Agent ID: Arkose Labs is introducing Agent ID, which follows the agent, not the IP: a stable identifier for each agent that holds when it rotates networks, ships a new build, or appears across many accounts.
• Introducing Detection Auditing: Arkose is introducing detection auditing, an audit trail for agent detection: when signals were updated, what each session was classified as at the time, and what a new signal would have caught, exportable from Command Center.
• Global Intel: How a business’s agentic traffic compares with the Arkose network: agent mix over time, attack campaigns and origin countries.
• Proof of Human (Coming Soon): For steps a business marks as sensitive, such as moving money, Arkose Labs is launching a way to require a person to confirm before an agent can continue.

Agent Trust Manager already shows businesses which agents in their traffic are self-disclosing good, non-disclosing good or malicious. The new controls let them act on that view by applying the right policy to each population of agents at each step.