The problem with voice and caller ID
A voice agent can say any name, and caller ID can be spoofed. A practice that wants to act on an agent's request has two bad options: refuse every automated caller, or accept claims it cannot check. Neither scales as more patients hand routine calls to assistants.
How the practice checks the caller with AgenID
The assistant presents its AgenID. The practice's agent resolves it and gets the operator's signed manifest, meaning who runs the agent and what it says it is for. It then re-verifies the signature itself. In the illustrated flow the practice also issues a fresh challenge that the agent answers with its key, so a recording of an earlier call can't be replayed.
The operator declaration shows as DECLARED at L1_REGISTERED, which is what the operator says about itself. A domain claim at L2_DOMAIN_VERIFIED would mean an authority checked that the operator controls the domain it names.
Identity is not access to records
Knowing which agent called does not give it access to anything. Whether it may see records, change a booking or act clinically stays entirely the practice's decision, under its own policies and the privacy rules it already follows. AgenID answers who; the practice decides what.
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 assistant book appointments for me?
Many can place the call. The open question is whether the business can tell which assistant is calling and who stands behind it. That is the gap a portable identity like AgenID fills.
How does a practice verify an AI caller?
It resolves the identifier the agent presents, re-verifies the operator's signature over the manifest, and checks which specific claims have been independently verified. It can do all of this without an account with AgenID.
Does verification give the agent access to patient records?
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.
What does the practice need to install?
Resolution is an HTTPS request plus standard Ed25519 signature verification. No adapter package is published yet for any scheduling or voice platform; the partner docs describe the integration pattern.