The IAB Tech Lab is filling in another piece of the infrastructure needed for agentic media buying.
The organization has introduced AAMP 3.0, the latest version of its Agentic Advertising Management Protocols framework (that’s a mouthful). The big addition is OpenProposal, a standard designed to let buyer and seller AI agents communicate during the traditionally manual stretch between an advertiser issuing an RFP and committing to a media buy.
Essentially, OpenProposal gives publishers a standard, machine-readable way to describe their ad products and respond to briefs. Buyer agents can then discover those opportunities, compare them against campaign requirements and determine which ones fit.
Why This Matters:
Agentic advertising is methodically moving from demos and experiments toward actual media buying. But getting agents to work within individual platforms is only part of the challenge, of course. For agentic buying to scale across the open ad ecosystem, agents built by different companies need common ways to communicate.
That’s the infrastructure problem AAMP 3.0 is trying to solve.
The RFP and proposal process has historically relied heavily on emails, documents, spreadsheets and other very human workflows. OpenProposal turns more of that process into structured information that agents can understand. A buyer agent could potentially evaluate opportunities across far more publishers than a human planner could reasonably consider manually.
Importantly, IAB Tech Lab isn’t proposing an entirely new transaction infrastructure. OpenProposal connects into existing standards including AdCOM, OpenDirect and the Deals API, creating a path from discovery and planning through negotiation, buying, execution and reporting.
Experts React:
In the announcement AAMP 3.0 launch release, Michael Palmer, global VP of emerging technologies at WPP Media, said AAMP 3.0 could help move the industry from “proof-of-concept tests towards the wide-spread use of AI Agents in media transactions.”
Also, here’s IAB Tech Lab CEO Tony Katsur on the launch:
Our Take:
The agentic advertising conversation has understandably focused on what AI agents can do: build plans, discover inventory, negotiate deals and execute campaigns. But the less exciting infrastructure underneath those capabilities will probably — in the end — determine if they can operate at scale.
Agents can’t efficiently transact across hundreds of companies if every buyer, seller and platform speaks a different technical language. Standards like AAMP and OpenProposal are an attempt to create that common language, and turn agentic from a hodgepodge of individual experiments into something that can work across the broader ecosystem.
