The advertising industry has been through protocol fragmentation before. Before OpenRTB became the standard for real-time bidding, each major exchange operated its own bid request format. A sell-side platform integrating multiple exchanges wrote different code for each, maintained separate connections, and absorbed the overhead of a fragmented interface landscape. The standardisation that finally simplified this was not inevitable. It required years of industry coordination, competitive pressure, and a de facto leader willing to publish and maintain a common specification. The current landscape for agentic advertising has two competing protocol efforts and no convergence date. The industry has seen this before. It knows how it ends.
TL;DR. The IAB Tech Lab's agentic advertising specifications and the Ad Context Protocol are currently developing in parallel, with overlapping scope and no announced path to convergence. If both achieve adoption without merging, buyers and sell-side platforms will face the same integration overhead that fragmented early programmatic: build twice, maintain twice, translate between formats when your counterparty uses the other one. The difference from early programmatic is that agentic deal records carry liability, not just bid parameters. A translation layer between two standards introduces the same semantic drift risk that a single shared record is designed to eliminate. The industry should be treating convergence as a technical necessity, not a competitive courtesy.
What happened in early programmatic, and why does it matter?
Before the IAB published OpenRTB 1.0 in 2010, real-time bidding ran on proprietary formats. A DSP integrating with multiple exchanges built a separate connector for each. The bid request from one exchange used different field names, different encoding, different required parameters than the bid request from another. The overhead was real: integration time, maintenance cost, and the ongoing risk that a field meant different things to different parties.
OpenRTB did not emerge because everyone agreed it was a good idea. It emerged because the fragmentation cost became large enough that the major players found common ground preferable to continued divergence. The standard was not perfect and required significant iteration, but it gave the industry a shared vocabulary for auction mechanics that, once established, proved difficult to displace.
The parallel is not exact. OpenRTB standardised bid request format: a relatively constrained set of parameters describing an impression opportunity. Agentic deal record standards are attempting to standardise the record of a bilateral negotiation: a more complex object with legal, financial, and operational implications beyond what a bid request carries.
Where do AAMP and AdCP currently agree and diverge?
Both bodies have converged on the same core requirement: agentic advertising requires a shared, machine-readable record of deal terms that both parties can reference and audit. Shailley Singh of IAB Tech Lab described this as "structured transaction objects, explicit transaction states, and systems of record that minimise hallucination or misinterpretation of context," as reported in an ADOTAT investigation published in August 2026. Brian O'Kelley of AdCP confirmed in the same investigation that a shared machine-readable record is a precondition for cross-company agentic buying.
The agreement on the requirement is complete. The agreement on what the record looks like, how it is transmitted, and who maintains it is not published. Both efforts are active; neither has announced a joint specification.
What does fragmentation look like in practice?
A publisher deploying an agentic sell-side platform would need to decide which protocol its agent speaks. If its largest DSP partners use AAMP and its largest agency holding company partners use AdCP, it either builds connectors to both or picks a side and excludes half the market.
A buyer's agent instructed to find inventory across a publisher mix would encounter sell-side agents speaking different protocols. Either the buyer builds translation capability, or the buy is limited to counterparties using the same standard. Programmatic solved this with OpenRTB. Without an equivalent convergence event, agentic advertising fragments along protocol lines.
The financial reconciliation risk is larger than in programmatic. A bid request translation error produces a failed bid. A deal record translation error produces a disputed deal. The semantic drift problem that a shared record is designed to eliminate reappears at the protocol boundary: if AAMP and AdCP structure the same deal information differently, translating between them is the same operation as asking two different language models to parse the same natural language into structured data. The divergence rate will not be zero.
What forces convergence, and how long does it take?
Historically, protocol convergence in advertising has followed one of two paths: regulatory mandate, or scale pressure from a dominant party. GDPR created convergence pressure on consent management frameworks. OpenRTB happened because the scale of the major exchanges made their format the de facto baseline and the IAB formalised it.
Neither path has activated for agentic deal records yet. The ADOTAT investigation confirmed that agentic buying is running at approximately $3,000 a day in gross media spend, as stated by Andrew Mole of pubX. At that volume, fragmentation is an inconvenience. At ten thousand times that volume, it is a structural cost. The window to establish convergence before scale makes it expensive is open now.
The standard body with the broader existing industry adoption -- the IAB Tech Lab, which maintains OpenRTB, the IAB's consent frameworks, and the existing auction infrastructure -- holds natural convening authority. AdCP's competitive differentiation is technical specificity and its lead author's track record. A joint working group between the two bodies is the mechanism the industry should be asking for, before the infrastructure for separate records calcifies.
________________
This article references the history of OpenRTB standardisation and the IAB Tech Lab's published agentic advertising specifications, statements by Shailley Singh of IAB Tech Lab and Brian O'Kelley of AdCP published in ADOTAT in August 2026, and volume data from Andrew Mole of pubX published in the same investigation.