AdCP: How the Ad Context Protocol Enables Agentic Media and AI-Driven Media Buying

Media is shifting from platform-led execution to protocol-led intelligence — a direction we’ve explored in pieces like Top 10 Agentic AI Plays Transforming Media. As AI agents evolve from assistants to autonomous planners and optimisers, they need a shared language — something the industry has never standardised.

Ad Context Protocol (AdCP) is the first serious attempt to change that.

It’s early. It’s experimental. It’s not widely adopted. But it’s earning attention from the architects shaping the next generation of ad infrastructure — and for anyone building future media workflows, it points clearly to what’s ahead.

From Human Translation to Agentic Execution

Today’s workflows still depend on operators manually bridging the gaps: hopping between platforms, rebuilding audiences, stitching spreadsheets, reconciling taxonomies, and re-entering everything by hand.

AdCP imagines a different world. A media operator — or their AI agent — could simply say:

Find high-intent sports audiences across retail media and CTV, compare CPMs, and activate the best option.

The agent interprets the intent.
The protocol standardises the request.
The system executes across platforms.

This intent-based workflow is aligned with the direction described in Agentic AI Spins the Retail Media Flywheel Faster.

It’s the bridge between fragmented operations and AI-native orchestration.

What AdCP Actually Standardises

AdCP extends the Model Context Protocol (MCP) into advertising, giving AI agents structured ways to describe campaigns, creatives, audiences, and performance.

It’s not real-time bidding.
It’s not retargeting.
It’s an interoperability layer.

Today it defines three core components, with a fourth emerging:

1. Creative Protocol

Consistent labelling and metadata for creative assets for clarity and alignment.

2. Media Buy Protocol

A standard set of tasks for discovery, planning, activation, and optimisation.

3. Signals Activation Protocol

Normalising how audiences and segments are represented across platforms.

4. Curation Protocol

A common framework for bundles, deals, and curated marketplace supply.

Together, these form the “grammar” AI agents use to operate across the media supply chain.

Transparency: The Feature That Changes the Game

AdCP is transparent by design — every request and response is structured, logged, and auditable. That creates a new dynamic:

Buyers gain clarity on spend, performance, and optimisation logic.
Publishers benefit from more direct, intelligent demand.
Intermediaries face a world where opacity becomes a disadvantage, not an asset.

This need for clarity mirrors our recent conversations Cutting Through the Noise – mediaNXT #Connect with Roger Dunn, where transparency and clean signals unlock better decision-making.

The transition won’t happen overnight — but the trajectory is unmistakable.

The MediaNXT View

At MediaNXT by Trigg, we see AdCP as a foundational layer in the shift to Agentic Media — where AI agents operate fluidly across Salesforce Agentforce Media, Data 360, media and retail media networks, and the wider ecosystem.

This isn’t about removing people from the process. It’s about removing the drag.

Less trafficking
Less spreadsheet stitching
Less rework

And in return:

More strategic focus.
More intelligence infused into every decision.

“Agentic Media doesn’t replace operators — it unleashes them. It gives teams a system that finally moves at the speed of their ambition.” Steven Paul, CEO, Trigg Digital

AdCP gives agents the structure to plan, buy, and optimise autonomously — while keeping humans firmly in charge of direction, governance, and outcomes.

The Bottom Line

AdCP isn’t the final answer — but it’s the clearest early signal of where media is heading.

As the ecosystem becomes more connected, more data-driven, and increasingly agent-enabled, the industry will move:

from platforms → to protocols
from manual activation → to autonomous orchestration
from black-box mechanics → to transparent, AI-interpretable standards

And when that shift lands, planners won’t be jumping between ten platforms.

They’ll just be instructing their agent.