Why per-platform accounts don't scale for agents
If every airline, hotel and venue issues its own agent account, identity fragments. The same agent is four strangers, its history can't follow it, and each platform carries the full cost of vetting it from scratch.
One AgenID, four independent checks
The travel agent presents the same AgenID everywhere. Each platform resolves it and re-verifies the operator's signature on its own, and none of them needs to trust another platform or a central account provider. The illustration shows each one confirming a domain claim at L2_DOMAIN_VERIFIED before releasing a seat, a room, a car or a table.
Because the identifier is permanent, the relationship each platform builds with the agent carries over to the next trip.
Bookings and payment stay with the platforms
Fares, room rates, cancellation rights and payment belong to each provider and its terms. AgenID only answers which agent is booking and who stands behind it.
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 book a whole trip?
Agents can already search and book across providers. What is missing is a way for each provider to know which agent it is dealing with, which is the gap a portable identity fills.
What does a portable AI agent identity mean?
It means one identifier and one signed record that any platform can resolve and verify, instead of a separate account per service.
Do platforms need to integrate with each other?
No. Each resolves the identity independently. Verification needs only the public record and standard Ed25519 signature checking.
Does verification entitle the agent to book?
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.