By Karsten Weide, Chief Analyst.
On October 15, 2025, a coalition of AdTech and AI advertising companies announced the Ad Context Protocol (AdCP)-an open standard meant to let AI agents communicate with advertising platforms and, over time, with each other. Founding members named include PubMatic, Optable, Scope3, Swivel, Triton Digital and Yahoo, with a broader roster of launch members: Accuweather, Adgent, Bidcliq, Butler/Till, Classify, HYPD, Kargo, Kiln, LG Ad Solutions, Locala, Magnite, Media.net, MiQ, Nativo, Newton Research, OpenAds, Raptive, Samba TV, Scribd, The Product Counsel, and The Weather Company. The group says the code is public on GitHub.
What Is AdCP – Today and Tomorrow?
Generally speaking: AdCP is positioned as an open, agent-first layer that standardizes how autonomous or semi-autonomous systems (think “AI media agents”) negotiate, plan, and execute advertising tasks across platforms. It’s described as parallel to OpenRTB -not a replacement- focused on orchestration and workflow, while OpenRTB continues to govern impression-level auctions. Practically, publishers and platforms could run OpenRTB and AdCP side-by-side.
What AdCP can do today: The initial scope centers on structured, interoperable “conversations” between agents and platforms – e.g., requesting inventory packages, exchanging campaign constraints, coordinating measurement, or triggering actions (create a line item; adjust frequency; pause a tactic). The standard is built on or aligned with modern agent frameworks (e.g., Model Context Protocol [MCP]) to make cross-platform agent messaging feasible.
Future perspective: The ambition is to automate larger chunks of the media lifecycle – from briefing and planning to pacing, optimization, and reconciliation – by letting agents “speak a common language” across DSPs, SSPs, ad servers, and data tools. If widely adopted, AdCP could let buyers’ and sellers’ agents coordinate multi-platform strategies without brittle, one-off integrations.
Will it just automate today’s programmatic ecosystem – or supersede it? Near-term, AdCP is adjacent to programmatic: it orchestrates workflows on top of existing pipes, rather than replacing auctions. Long-term, if agent-to-agent negotiation matures, parts of the market could shift from today’s impression-by-impression bidding toward goal-driven negotiations (budgets, outcomes, constraints) that wrap around OpenRTB. In other words, AdCP aims to abstract the messy plumbing, not immediately rewrite it.
State of Adoption and Prospects
AdCP is brand-new. The spec and code are public; a couple dozen companies have attached their names; trade coverage frames this as groundwork for agentic advertising, not a completed standard. The breadth of launch signatories (sell-side, data, measurement, and some buy-side services) suggests credible momentum, but no large, universal deployment yet. Adoption prospects hinge on whether AdCP quickly ships usable reference adapters and clear governance so that joining is low-risk for platforms already stretched by privacy, identity, and CTV complexity.
Challenges and Criticisms
- Governance & neutrality: Who steers the roadmap, certifies compliance, and resolves disputes? Without robust, independent governance, skeptical players will see “open” as someone else’s agenda.
- Overlapping standards: Markets already support OpenRTB, OpenDirect, Prebid, UID frameworks, and seller-defined signals. Teams may ask, “One more protocol?” AdCP must show net simplification of integrations, not added burden.
- Security & liability: Agentic systems that can take actions (spend, pause, reallocate) demand hardened authentication, audit, and rollback. Standards will need granular permissions and transparent logs. (Advocates say this is planned; it still must be proven in production.)
- Latency and interoperability: Real-time control loops across multiple platforms and agents can introduce coordination lag unless implementations are efficient and consistent.
- Competing incentives: Large incumbents may prefer their own orchestration layers, limiting universal adoption. (See below on The Trade Desk and Google.)
Strategic Consequences if AdCP Scales
For DSPs: AdCP could commoditize peripheral workflows (e.g., trafficking, reporting glue) while shining a spotlight on core differentiation-identity, optimization, unique demand, and partner access. Best-in-class DSPs might gain if agents can more easily test multiple platforms; weaker ones lose the “integration moat.” Conversely, if agents can call platform features directly, DSP UIs become less sticky, but algorithms and inventory pipes matter more.
For SSPs and publishers: Standardized agent requests could reduce RFP friction, expose real buyer preferences, and make package and deal management more dynamic. In CTV and premium video, seller agents could proactively assemble goal-aligned deals. If AdCP thrives, sell-side data and packaging become higher-margin, while pure pass-through exchange functions face fee pressure.
For agencies: Agentic workflows will pull agencies toward strategy, brand safety policy, and measurement frameworks while automating repetitive trafficking and pacing. Agencies that codify their playbooks into agents (budget rules, suitability, incrementality tests) could scale their “secret sauce” across platforms more quickly.
For advertisers: If AdCP delivers, brands get faster experimentation, cleaner cross-platform controls, and more comparable reporting. Financially, that can translate into lower ops cost and better marginal ROI from continuous, agent-driven optimization – if verification and governance keep pace.
Why Are The Trade Desk and Google Not Founding Members?
What explains the conspicuous absence of The Trade Desk and Google among the founding members?
- Strategic alternatives in flight: The Trade Desk has just rolled out OpenAds and related sell-side tooling-its own vision for a fairer, transparent auction and a clearer buyer-seller handshake-which may reduce its appetite to back a new external control layer right now. Google tends to prioritize its proprietary end-to-end stack and historically supports cross-industry standards selectively.
- Control and pacing: Both companies have strong incentives to shape any emerging “agent layer” to protect core moats (optimization, identity graphs, auction mechanics). Joining at launch could constrain their options.
Bottom Line
AdCP is an ambitious, agent-first coordination layer that aims to make cross-platform planning and buying less brittle and more automated. It does not replace OpenRTB; it wraps around the existing programmatic pipes to standardize how agents talk to platforms. Success will depend on real, open implementations, credible governance, and early proofs that AdCP reduces integration cost while improving outcomes. If that happens, the financial and strategic center of gravity could shift toward agent-defined goals and platform-agnostic execution-rewarding those vendors (buy- or sell-side) with the strongest optimization, identity, and supply quality, and squeezing those whose value was mainly workflow glue.
Leave a Reply