Know Your Agent.
Also known as: KYA
Know Your Agent is an emerging approach to checking an AI agent before allowing it to act. Who is it acting for? What has it been given permission to do? And is the action it is trying to take within those limits? Until recently, it was usually a person clicking through a website, filling in a form, or making a purchase. Now it could be an agent doing some of that on their behalf. The name is borrowed from Know Your Customer, or KYC, the identity checks used in financial services and other industries. KYA is a developing practice, rather than one settled global legal standard. Businesses and standards bodies are working on ways to verify agents and the authority they have been given. Existing laws and regulations can still apply to what an agent does.
In the wild
“Instead of asking “Who is the customer?”, KYA asks a broader question: Who is this agent, who is it acting for and is it authorized to act?”
Seen here: Experian, June 2026
For AI users
Know what you have allowed your agent to do, what it can access, and whether it can spend money, share information, or make commitments without asking you. Check how to change those permissions or switch them off. You may not see every check happening. A business could verify your agent’s identity and permissions in the background before allowing it to act for you.
For business leaders
As customers start sending agents to interact with your business, you need to know whether the agent is legitimate and whether the customer actually wanted the action or transaction to happen. Checking identity, permissions, and the customer’s instructions could help reduce fraud and give everyone a clearer record if a transaction is disputed. This is a conversation for boards to be having now, even though there is no single settled KYA framework to follow. Ask who owns agent identity, how permissions will be checked, what records you will keep, and what happens when something goes wrong.