An open standard that lets AI agents talk directly to ad platforms to find inventory, check pricing, and place a buy. For agencies, it is the wiring that decides whether an agent can actually execute a campaign or just describe one.
Also known as AdCP
Ad Context Protocol, or AdCP, is an open standard built on top of Anthropic’s Model Context Protocol that gives AI agents a common language for talking to demand-side platforms, supply-side platforms, and publishers. Instead of every ad tech vendor building its own one-off integration for every AI agent that wants to buy media, AdCP gives agents one consistent way to ask what inventory is available, what it costs, and how to activate it.
The protocol launched in October 2025 with backing from more than 20 companies including PubMatic, Yahoo, Magnite, and Scope3, and it is governed by a neutral working group rather than owned by a single vendor. It currently covers three areas: audience activation, curation, and media buying.
AdCP is the plumbing underneath a trend agencies cannot opt out of: clients and platforms are already routing real budget through agents.
It changes who agencies compete with. Omnicom has confirmed live client media buys executed through an agent-to-agent framework built on AdCP, which means the agency’s edge shifts from manual trading skill to how well it configures and governs the agents doing the trading.
It sets a floor for transparency. Because AdCP standardizes how pricing and inventory get surfaced, agencies lose some of the opacity that used to justify certain markups, and clients will start asking why a human needed to be in that loop at all.
It creates a new vendor question. Any DSP, SSP, or publisher an agency works with either supports AdCP or doesn’t, and that support, or its absence, is becoming a real factor in platform selection, not a footnote.
A mid-size agency running programmatic for a retail client sets up an AdCP-compatible trading agent to handle audience discovery and initial bid requests across three DSPs overnight. The agent surfaces inventory and pricing by 7am, flags two placements that violate the client’s brand safety rules, and drafts a buy recommendation. A media strategist reviews the recommendation over coffee, adjusts the budget split between two publishers, and approves the buy before 9am, work that used to take a trader most of a morning.
The automations and agents module of the workshop teaches you how to build AI workflows that compress the busywork without taking the craft out of the studio.