IAB Tech Lab wants agents to review your RFP responses. Its new OpenProposal specification is being sold as a fix for a bandwidth problem: publishers miss briefs they never see, buyers miss inventory they never had time to evaluate, and agents can finally read all of it. I do not doubt that part works. What I doubt is who actually wins when the reading is automated.
The case for OpenProposal, stated fully
The argument in favor is real and I want to give it its full weight before I argue against it. Today, a media team might have time to evaluate proposals from ten or twenty publishers for a single campaign while hundreds of others never see the brief at all, simply because a human cannot read that many decks. IAB Tech Lab’s own framing puts it plainly: “Agents can review far more options than a human team ever could, which is what makes them so powerful.” OpenProposal standardizes eleven categories of proposal data, from audience reach to commercial terms to prohibited adjacency, so an agent can assess a small publisher’s response on the same terms as a familiar large one. For publishers who have never had a relationship with a given buyer, that is a genuine door that was previously closed. For agencies, it is less time reconciling spreadsheets and more time on strategy. I am not arguing that this is nothing. It is a real efficiency gain, and pretending otherwise would be dishonest.
But the gate does not disappear. It moves.
Here is my objection: OpenProposal changes who reads the proposal, not who controls the reading. Before, the gatekeeper was a media planner’s limited time and a rolodex of familiar publishers. After, the gatekeeper is whichever agent the buyer runs, tuned by whichever model and whatever criteria that buyer’s platform bakes in by default. A format that machines can parse is not the same thing as a marketplace that treats every proposal fairly. IAB Tech Lab’s own description of the specification calls it a foundation for agentic workflows to build on, not a finished, audited marketplace, and that distinction matters more than the press language around it suggests.
The publishers most likely to benefit from being newly “discoverable” to agents are the ones whose data is already clean enough for an agent to parse confidently. That favors large, well-resourced sales teams who can invest in structuring their proposals to whatever schema wins adoption, which is close to the opposite of the small or new publisher OpenProposal is pitched as helping. Scale advantages do not vanish when you hand the reading to software. They just move up a layer, from who has the best relationships to who has the best-formatted data pipeline.
What it means for the marketing leader
None of this is a reason to sit out standardization. A shared format for proposals is better than the status quo of a dozen incompatible ones, and IAB Tech Lab’s track record on transparency and consent frameworks is not nothing. But a CMO or head of programmatic evaluating any “agentic RFP” pitch, from OpenProposal or a vendor tool built on it, should ask a sharper question than “how much time does this save.” Ask which publishers actually got shortlisted before and after, not just how many proposals the agent reviewed. Ask whether the agent’s ranking criteria are disclosed and auditable, or a black box the platform vendor controls. Efficiency claims are easy to verify. Fairness claims are not, and the industry has a habit of only checking the former.
Ibexa’s own launch of its Agentic Marketing Platform this same week, covered here, makes the same “agents act, humans still decide” promise on the execution side of the stack. That promise is only as good as the guardrails a buyer actually inspects, not the ones a vendor ships by default. Demandbase’s Mojo launch is a third recent example of the same trade-off: more gets automated, and the question of who set the rules the automation runs on gets less visible, not more.
Standardizing the language proposals speak is a genuine improvement. Pretending that improvement is neutral to who holds power in the transaction is not honest, and it is worth saying so before the industry spends the next year treating “agentic” as a synonym for “fair.”
Source: IAB Tech Lab
