Yahoo, PubMatic, and Others Launch New Standard for Agentic Advertising

More than 20 adtech firms are supporting the Ad Context Protocol to cut costs, boost transparency, and lay the groundwork for agentic advertising.

A consortium of ad tech companies is laying the groundwork for the next evolution of automated advertising. More than 20 companies including PubMatic, Optable, Scope3, Swivel, Brian O'Kelley-backed Classify, Triton Digital, Magnite, Kargo, and Yahoo have signed on to the Ad Context Protocol (AdCP), an open technical protocol to standardize how AI agents - which carry out tasks on behalf of humans - communicate with one another. The standard aims to make this new technology more consistent across advertising platforms and allow for interoperability. The code for AdCP went live Wednesday on GitHub and is open to the public.

The initiative mirrors how OpenRTB standardized programmatic advertising more than a decade ago, said Kyle Dozeman, chief revenue officer at PubMatic. Backers say AdCP could do the same for the emerging era of agentic advertising, in which AI agents plan, buy, and sell media without human intermediaries.

AdCP was developed by a consortium of publishers, ad tech firms, and data providers, and is governed by a neutral working group rather than a single vendor. The protocol is built on Anthropic's Model Context Protocol, a framework that lets AI systems communicate using shared context, and provides an interface for AI agents to discover audiences, negotiate media buys, and execute campaigns across platforms.

The effort comes as both walled gardens like Meta and independent ad tech firms rework their infrastructure to accommodate AI agents. Early backers say AdCP could reduce costs, improve transparency, and pave the way for a new generation of AI-native advertising systems.

Bypassing programmatic

At launch, AdCP includes two main tools: one that helps advertisers find and reach the right audiences (Audience Activation Protocol), and another that lets AI agents buy and manage ad placements automatically (Media Buy Protocol).

In practice, a brand could ask its AI agent to "find women interested in rock climbing in the U.K." Publisher and platform agents could respond with matching audiences and ad opportunities. Once a deal is agreed upon, the transaction is written directly into the ad server - bypassing traditional programmatic auctions and enabling richer, custom ad formats, according to Anne Coghlan, co-founder and chief operating officer at Scope3.

“The benefit there is that the advertiser who has their brand agent doesn’t need to have individual integrations with every single one of the publishers in the world,” she said.

Opening up premium inventory

AdCP has helped publishers like The Weather Channel and weather.com to build new selling agents, presenting inventory, audience, and data assets without ceding control to opaque supply chains, according to Dave Olesnevich, head of product at The Weather Company.

"It opens up the opportunity to harness the full value of our first-party data," Olesnevich said. "We're seeing more control over how we represent our premium inventory, which often gets lumped into generic impressions or blocked by brand safety tools."

The launch underscores how ad tech firms are beginning to collaborate on shared infrastructure for agentic advertising — a shift that could reshape long-standing dynamics between publishers, platforms, and advertisers as AI agents take on more of the media planning and buying process.