Explore / Governance, Risk & Compliance
Know Your Agent (KYA)
If an AI agent can act for your business, you need to know what it is allowed to do.
Practical guidance for business owners and teams reviewing an agent’s access, authority and oversight. Know Your Agent (KYA) is an emerging term for establishing an AI agent’s identity, who is accountable for it, what authority it has, and how its actions are controlled. KYA is not one universal legal standard or a certification that automatically makes an AI agent safe.
Six checks behind a well-controlled AI agent
- Identity: Which agent is taking the action?
- Ownership: Who is accountable for it?
- Permissions: What can it access?
- Authority: What may it actually do?
- Evidence: Can we review what happened?
- Control: How do we stop or change it?
A practical KYA readiness checklist
- Define the purpose, deployed agents, suppliers, and accountable owner.
- Document systems, data, tools, access limits, and human approval points.
- Test expected tasks, errors, misuse, and attempts to exceed authority.
- Keep proportionate evidence, monitor for unexpected activity, and test how to revoke access.
How KYA fits with AI governance and security
KYA does not replace identity and access management, cybersecurity, data protection, or AI governance. It is a way to check that established practices remain effective when software takes actions on behalf of people or organisations.
This page provides general educational guidance. It is not legal advice, a security assurance report, or a certification. Controls should be assessed against the actual agent, connected systems, business context, and applicable laws and contracts.