Ad Industry Moves to Set Rules Before AI Agents Control Media Budgets

A new initiative called the Ad Context Protocol, or AdCP, is emerging as a standardized way for AI agents to communicate with publishers, ad platforms and each other. Backed by companies including Yahoo, PubMatic, Scope3, Optable, Swivel and Triton Digital, the protocol is designed to let software agents plan, negotiate and activate campaigns through a common machine-readable interface rather than a patchwork of proprietary APIs, according to coverage in Adweek and Digiday.

Protocol for Machine-to-Machine Buying

Adweek described AdCP as an “agent-first” media buying standard backed by more than 20 companies seeking to create interoperability across the ecosystem, while Digiday reported that the effort is intended to operate alongside, not replace, existing pipes such as OpenRTB.

Today’s programmatic ecosystem relies on auction standards and exchange integrations, but most workflows still assume human operators navigating dashboards. AdCP is different. It defines a schema for expressing campaign intent, audience parameters, pricing structures and performance goals in a way agents can directly interpret and execute against. This protocol outlines modules for media discovery, creative generation and signal activation, structured for machine-to-machine communication.

The project is open source and MIT licensed, signaling an attempt to attract broad participation and avoid vendor lock-in. That governance choice is deliberate. As Scope3 argued in explaining why AdCP matters, interoperability will be critical as AI agents begin transforming advertising workflows.

Why Infrastructure Is Being Built

The timing reflects a broader shift underway across digital commerce. As autonomous systems gain the ability to transact on behalf of users or enterprises, industries are moving to define protocol layers before agent activity scales.

In payments, Visa has introduced developer tools to support agent-driven commerce, positioning itself as infrastructure for autonomous transactions, as previously reported by PYMNTS. Google, meanwhile, has promoted Model Context Protocol servers as a way to standardize how AI systems retrieve verified data across environments, according to PYMNTS coverage of its Data Commons MCP server.

Advertising appears to be following that pattern. Adweek reported that industry executives increasingly see agentic AI as inevitable in campaign execution, prompting early moves toward shared standards. The concern is straightforward: If AI agents begin allocating meaningful portions of media budgets, fragmentation at the integration layer could stall adoption or concentrate power among platforms with closed ecosystems.

Executives involved in the AdCP effort describe it as additive rather than disruptive to existing pipelines. It can run alongside OpenRTB and other standards, allowing gradual adoption, Digiday noted. But the strategic calculation is clear. If agents are going to negotiate pricing, select inventory and optimize in real time, those interactions need structure, auditability and shared definitions.

Without that, buyers would be forced to build and maintain separate integrations for each publisher and platform, undermining the efficiency gains autonomous systems promise.

The Real Stakes

If AdCP gains adoption, it could lower switching costs for advertisers by making inventory more interoperable and comparable across publishers. It could also create new monetization pathways by allowing publishers to package inventory in formats designed specifically for machine consumption. Scope3’s analysis argues that interoperability will be essential if agents are to execute across multiple systems without friction.

But shared protocols also redistribute influence. The entities that define governance, extensions and updates to the standard may wield outsized power over how autonomous buying evolves. Competing frameworks are emerging, suggesting consensus is far from settled, as Adweek has reported in broader coverage of agentic AI in advertising.

For now, with AdCP, adoption is limited, and most ad budgets remain human-directed. Yet the industry’s move to organize around a protocol signals a recognition that AI agents are becoming potential economic participants.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.