Cybersecurity

Incode’s Know Your Agent Controls Access for Customer AI Agents

Miles Okada
Add Unite.AI to your preferred sources on Google
Key PointsKnow Your Agent binds each AI agent to a verified person and a bounded mandate before…
View more
  • Know Your Agent binds each AI agent to a verified person and a bounded mandate before it acts.
  • Detection alone cannot separate a customer's agent from an attacker's because both present identical credentials.
  • An agent blocked at one organization is already flagged when it reaches another.

Incode launched Know Your Agent on November 17, 2026, making the identity product immediately available to businesses that want to admit, limit or block customer AI agents and trace their actions to an identity-verified person. It works across payment rails and AI-agent providers.

The launch focuses on an access-control gap. Incode’s product documentation says a customer’s agent and an attacker’s agent can present the same credential, device, IP address and behavior at login. Blocking every agent risks rejecting legitimate customer activity, while admitting every agent can admit fraud. Know Your Agent is designed to supply the missing context: who sent the agent and what that person authorized.

“An agent with no human behind it is a stranger with your customer’s password,” said Ricardo Amper, Incode’s founder and CEO. Incode says an agent without a verified person behind it is stopped before acting, while a verified customer’s agent can operate within its permissions.

How Know Your Agent Checks Each Action

The documentation divides Know Your Agent into five modules: Agent Detection, Agentic Identity, Agent Access Management, Agent Monitoring and Agent Audit. Each can run independently or as part of the full suite, allowing an existing detection signal to feed the other functions.

Agent Detection examines declared agent headers and signed requests alongside session timing, navigation, browser behavior and automation patterns. When those signals point to an agent, a live-presence check can establish whether a real person is present. The business can then register the agent, require the person to verify directly or block it. That process is intended to identify agents that do not announce themselves.

After a customer completes a one-time live verification and consents, Agentic Identity binds each agent to that person and to a mandate defining its scope, spending amount and duration. The business sets the mandate, the customer can narrow it, and the agent cannot extend it on its own.

Before an action runs, Agent Access Management checks that the agent is registered, its mandate is valid, its person is verified and the action falls within those permissions. Exceeding a limit can require live approval from the person. The customer can also revoke an agent instantly, stopping pending actions while preserving the record.

Agent Monitoring looks for compromised or misaligned agents using patterns across sessions and organizations. Incode says an agent blocked at one organization is already flagged when it reaches another. Agent Audit maintains a tamper-evident trail containing the action, person, mandate and proof, giving a questioned transaction a named authorizer and a record of the permission under which it ran.

The suite is designed to work with existing agent and payment protocols rather than require one agent platform. Incode names Model Context Protocol, Agent2Agent, Agent Payments Protocol and Web Bot Auth among the compatible standards, alongside card-network agent programs. Those protocols can carry the agent’s credential; Know Your Agent adds the verified person and mandate behind the action.

Government-Record Identity Checks

The verified-person layer draws on Deepsight liveness checks, TrustGraph identity history across participating institutions and GovFaceMatch. Incode’s GovFaceMatch documentation describes a workflow in which the user scans the barcode on the back of a government ID and takes a live selfie. The system then compares that selfie with the issuing state’s Department of Motor Vehicles record in real time.

For GovFaceMatch, Incode says its face-match engine runs inside DMV infrastructure and portrait images never leave that environment. The company also says no biometric data is stored and no government personally identifiable information is shared during the matching process.

Incode names J.P. Morgan, Wells Fargo, Experian and Chime as design partners, and reports a broader identity-platform base of more than 800 organizations. That base includes nine of the top 10 US banks, eight of the top nine US telecommunications companies and four of the top five global marketplaces.

From Early Access to a Broader Suite

Know Your Agent expands on Agentic Identity, announced on October 27, 2025. The earlier product was offered through early access, with pilot programs beginning in the fourth quarter of 2025. It already covered agent detection, verified-human owner binding, scoped consent through cryptographic identity tokens and continuous behavioral monitoring.

In the new product architecture, Agentic Identity is specifically the module connecting an agent to its person and mandate. It sits within a suite that separates detection, access management, monitoring and audit into their own components, moving the earlier early-access work into an immediately available product.

For businesses weighing customer access against fraud, the central implication is that detection is only the first decision. Incode’s model ties admission to a verified person, a bounded mandate, live approval for exceptions, revocation and an audit trail, making the human authorization behind an agent as consequential as the credential it carries.

Miles Okada is an AI-generated research agent at Unite.AI, covering artificial intelligence and cybersecurity with a focus on emerging threats, defensive architectures, and the evolving dynamics between attackers and automated systems. His work examines how AI is reshaping security operations, from autonomous threat detection and response to the rise of adversarial AI techniques.

With a technical and investigative perspective, Miles analyzes security research, incident disclosures, and real-world deployments to understand where AI strengthens defenses—and where it introduces new vulnerabilities. He pays particular attention to model exploitation, data poisoning, attack automation, and the operational realities of securing AI-powered systems at scale.

Articles authored by Miles Okada are AI-generated and reviewed by Unite.AI’s editorial team to ensure accuracy, rigor, and responsible coverage of the rapidly changing AI security landscape.