16 Nov 2026 · 3 min read
Agent Identity: How Do You Know the Agent You're Talking to Is Who It Claims to Be?
TL;DR: An agent identifying itself as representing a specific buyer or publisher is a claim, not proof. Before a deal is struck, each side needs a way to check that claim against something more reliable than the agent's own say-so. This is what agent identity verification is for, and it is a distinct problem from checking whether the deal terms themselves are fair.
Why a claimed identity isn't enough
When a buying agent contacts a selling agent, or vice versa, the first thing each side implicitly assumes is that the other agent genuinely represents the organisation it claims to. Without some way to check that, a selling agent has no real way of knowing whether it is negotiating with an authorised representative of a real advertiser or with something else entirely, a misconfigured test system, a scraper, or a deliberately misrepresented actor. The claim an agent makes about who it represents is only as trustworthy as the mechanism checking it.
What identity verification is actually checking
Identity verification in this context answers a narrower question than it might sound like: not whether an agent is trustworthy in general, but whether it is authorised to act on behalf of the specific organisation it claims to represent, for this specific kind of transaction. IAB Tech Lab's AAMP framework addresses this directly, treating identity checks as a defined part of how an agentic transaction is set up, separate from the negotiation itself. An agent passing identity verification tells a counterparty "this agent is authorised to represent this organisation." It does not tell them anything about whether the terms that agent then proposes are reasonable.
Why this is a separate problem from deal governance
It is worth being precise about the boundary here. Identity verification happens before or at the start of a transaction and answers who the agent represents. A deal record, such as Alkimi's DealSheet, exists after terms are agreed and answers what was actually committed to. A system could verify identity perfectly and still have no record of what the two verified agents actually agreed, and a system could hold a perfect record of agreed terms between two agents whose identity was never properly checked in the first place. Both problems need solving, and solving one does not solve the other.
Frequently asked questions
Who is responsible for checking an agent's identity, the buyer's side or the seller's side?
Typically both sides have an interest in verifying the other, since either side could otherwise be transacting with an unauthorised or misrepresented counterparty. Standards like AAMP are built to give both sides a shared way to do this rather than relying on one party's internal checks alone.
Does identity verification prevent an authorised agent from making a bad deal?
No. An agent can be fully verified as representing a real organisation and still propose or accept poor terms, exceed its mandate, or make an error. Identity verification is about who is at the table, not about what happens once they start negotiating.
Is there a certification that proves an agent is verified?
There is no universal compliance badge across the industry at this stage. IAB Tech Lab runs an Agent Registry where companies can list agents, but a listing is not the same as a guarantee of behaviour, and the specific verification mechanics vary depending on which standard and platform are involved.