Skip to main content

Documentation Index

Fetch the complete documentation index at: https://agenticadvertisingorg-changeset-release-main.mintlify.app/llms.txt

Use this file to discover all available pages before exploring further.

A3: The AdCP landscape

Free module — No account required. ~15 minutes with Addie. Prerequisite: A2.
This is the survey course. Touch everything, go deep on nothing. You’ll get a map of the entire AdCP ecosystem — every protocol domain, every discovery mechanism, every governance layer. Designed to pique your curiosity and show you where to go deeper. Completing A1 + A2 + A3 earns the AdCP basics credential.

Learning objectives

  • Describe what brand.json is and why every brand should have one
  • Explain adagents.json and how agents discover each other
  • Name all eight AdCP protocol domains and what each covers
  • Identify which areas you want to explore deeper in a role track
  • Recognize the three trust primitives that make the ecosystem safe: authenticated identity, signed governance tokens, and agent/account isolation

Reading list

Brand protocol

Brand identity claims, brand.json discovery, and brand hierarchy. Four variants: house portfolio, brand agent, redirects, minimal.

Governance overview

The governance protocol: property governance, brand governance, content standards, creative governance, and campaign governance.

Campaign governance

Always-on compliance: ties campaigns to media plans, validates every transaction through three independent parties.

Policy registry

Community-maintained library of advertising regulations and standards that brands reference by ID.

Embedded Human Judgment

The five principles behind AdCP governance — how humans stay in control when agents act autonomously.

Creative protocol

The creative protocol: assets, formats, manifests, creative agents, and 20 channels of adaptation.

Signals overview

The signals protocol: audience segments, contextual signals, measurement data, and optimization.

Sponsored Intelligence

Conversational brand experiences in AI assistants — a genuinely new advertising model.

Trusted Match Protocol

Real-time execution: context match and identity match activate packages at serve time across all surfaces.

Capability discovery

How agents advertise what they support so other agents can discover them.

Accounts protocol

The commercial layer: advertisers, operators, authentication, billing, and account lifecycle.

Security Model

How AdCP defends against unauthorized spend, cross-tenant leakage, replay attacks, and SSRF across the ecosystem.

What you’ll cover with Addie

Addie walks through each domain with a quick live example — just enough to understand what each area does: Discovery and community:
  • brand.json: your brand’s machine-readable identity at /.well-known/brand.json
  • adagents.json: how publishers declare which agents can access their inventory
  • Community registry: how agents and brands find each other
  • AgenticAdvertising.org: working groups, industry councils, how the spec evolves
Trust primitives (how the ecosystem stays safe):
  • Authenticated identity — cryptographic proof of which agent is calling, not a header claim the seller has to take on faith
  • Signed governance context — approval tokens from the buyer’s governance agent that bind a plan to a specific seller and phase, verifiable by auditors years later
  • Agent and account isolation — every piece of state is scoped to the calling agent and the account it is acting on; cross-tenant reads are rejected with a generic “not found”
Protocol domains:
  • Accounts — commercial identity, operator-billed vs agent-billed, account lifecycle, get_adcp_capabilities
  • Media buy — proposals, forecasting, refinement, packages, keyword targeting, geo-proximity
  • Creative — formats vs manifests, 20 channels, AI-powered generation with build_creative
  • Signals — audience data, privacy-compliant signals, conversion tracking, attribution
  • Governance — content standards, property lists, the Oracle model for AI-driven brand safety, campaign governance (multi-party validation, budget authority, policy registry)
  • Sponsored Intelligence — conversational brand experiences, cost-per-conversation instead of CPM (experimental in 3.0)
  • Trusted Match Protocol — impression-time execution: context match (content fit) and identity match (user eligibility), cross-publisher frequency capping, works across web, mobile, CTV, AI assistants, and retail media
  • Brand Protocolbrand.json resolution, brand identity claims, brand hierarchy
  • Registry — entity resolution, agent discovery, community directory

Assessment

DimensionWeightWhat Addie looks for
Breadth35%Can you describe what each protocol domain does?
Discovery mechanisms25%Do you understand brand.json, adagents.json, and capability discovery?
Key concepts25%Can you explain format vs manifest, billing models, the Oracle model, and context match vs identity match (TMP’s two-operation model)?
Synthesis15%Can you connect concepts across domains without prompting?
Passing threshold: 70%.

Start this module

Start A3 with Addie

Open Addie and say “I’d like to start certification module A3.”
Next: Choose your role track — Publisher, Buyer, or Platform