# Know which agent is acting and why.

Source URL: https://www.affix-io.com/agent-verification/

Summary: AI agent verification checks agent identity, delegated authority, wallet session scope, policy fit and audit evidence before an autonomous action is allowed.

Primary answer: An AI agent is verified when the system can show which agent is acting, who or what authorised it, what scope applies, and whether the current action fits policy. AffixIO turns that check into a signed yes or no decision.

Key facts:
- AffixIO returns a signed allow or deny decision for verification workflows.
- Source records, consent records and account data stay with the customer.
- Evidence can include proof_id, policy version, ML-DSA-65 attestation and Merkle audit material.
- AffixIO is a verification and attestation layer, not a payment network, issuer processor or identity warehouse.

Questions answered:
- What does it mean for an AI agent to be verified?
- What identifiers matter?
- Can this work without exposing PII?