Why support desks are wary of agent requests
Account changes are a common route for takeover and abuse, so support teams follow strict scripts. An agent that cannot be identified either gets stuck in those scripts or gets refused outright. Neither helps the customer who delegated a routine task.
How support checks the personal agent
The personal agent presents its AgenID. The support agent resolves it and re-verifies the operator's signature on the manifest. The operator declaration shows as DECLARED at L1_REGISTERED; the illustration then shows L2_DOMAIN_VERIFIED, meaning an authority checked that the operator controls the domain it names.
Because the identifier is persistent, the provider can see the same agent return for the next request, and the confirmation goes back to the same identity that asked.
The account holder is still the provider's to confirm
Identifying the agent is not the same as authenticating the account holder. The provider still confirms that the request comes from its customer, applies its cancellation terms and decides any refund, using the controls it already runs.
Today the reference registry at agenid.com issues DECLARED and L1_REGISTERED. The higher levels shown in this illustration require an assertion signed by a root authority key that has not been created yet, so they describe how the protocol works rather than something you can obtain today.
Frequently asked questions
Can an AI agent cancel subscriptions for me?
Agents can already contact support and make requests. What the provider usually cannot tell is which agent is asking and who stands behind it, which a portable identity makes checkable.
How does a support team verify an AI agent?
It resolves the agent's AgenID, re-verifies the operator's signature over the manifest, and reads which claims have been independently verified, without an account with AgenID.
Does verification let the agent change my account?
No. AgenID answers which agent is acting and who declared it. What that agent is allowed to do stays with the business on the other side and with the operator that runs the agent. Protocol v1.1.1 defines no signed authorization object.
Does AgenID hold my account credentials?
No. AgenID never holds account passwords, payment details or an operator's signing key. The signing key is generated in the browser at /issue and never sent to AgenID.