Insights / Ecosystem

Understanding the Agentic Protocol Landscape

How horizontal middleware and vertical standards work together

Overview

As AI agents become more capable, they need standardized ways to communicate with each other and with business systems. The emerging agentic protocol landscape addresses this through two complementary approaches: horizontal middleware that provides foundational agent-to-agent communication, and vertical protocols that enable specific business transactions.

Why Standards Matter

Without open standards, the AI agent ecosystem risks fragmenting into proprietary silos where agents from different vendors cannot interoperate. Open standards enable:

The Horizontal/Vertical Framework

The agentic protocol landscape can be understood through two layers:

Horizontal Middleware
Agentic AI Foundation (AAIF)
MCP • A2A • Agents.md • Goose
Vertical Transaction Protocols
AdCP
Advertising
ACP
Commerce
Future
Service, etc.
Traditional Infrastructure
IAB Tech Lab • OpenRTB • VAST • etc.

Horizontal middleware provides foundational agent communication. Vertical protocols enable specific business domains. Traditional standards remain the underlying infrastructure.

Key Insight

Horizontal middleware (AAIF) and vertical protocols (AdCP, ACP) are complementary, not competing. AAIF focuses on core agent infrastructure, while vertical protocols focus on domain-specific transactions. They work together to enable the agentic future.

Horizontal Middleware

Horizontal middleware provides the foundational layer for agent-to-agent communication. It defines how agents discover each other, exchange messages, share context, and coordinate actions - regardless of what specific business domain they operate in.

Agentic AI Foundation (AAIF)

Linux Foundation Project

The Agentic AI Foundation is a vendor-neutral home for open source agentic AI projects hosted by the Linux Foundation. AAIF provides open governance, funding for community programs, and coordination to ensure standards evolve for the community's benefit rather than any single company's interests.

Founding projects contributed to AAIF:

  • Model Context Protocol (MCP) - Anthropic's protocol for connecting AI assistants to external tools and data sources
  • Agent2Agent (A2A) - Google's protocol for communication between AI agents
  • Agents.md - OpenAI's specification for agent capability documentation, adopted by 60,000+ projects
  • Goose - Block's open source, local-first AI agent framework

Members include: Amazon Web Services, Anthropic, Block, Bloomberg, Cloudflare, Google, Microsoft, OpenAI, Shopify, Snowflake, and many others.

Visit aaif.io

What AAIF Focuses On

AAIF aims to be "what the W3C is for the Web" - a set of standards and protocols that guarantee interoperability, open access, and freedom of choice. Its focus areas include:

AAIF focuses on the core infrastructure - the "plumbing" that makes agent communication possible. It intentionally does not focus on domain-specific business logic, which is where vertical protocols come in.

Vertical Transaction Protocols

While horizontal middleware handles agent communication fundamentals, vertical protocols define how agents conduct specific types of business transactions. These protocols build on the foundational layer to enable domain-specific workflows.

Ad Context Protocol (AdCP)

Governed by AgenticAdvertising.org

AdCP is an open standard for AI-powered advertising workflows. It enables agents to express advertising intent, discover inventory, exchange signals, and execute media buys across publishers and platforms through a unified interface.

Key capabilities:

  • Natural language inventory discovery and brief-based targeting
  • Signal activation for audiences, contexts, and locations
  • Media buy creation and optimization
  • Creative format discovery and asset management

AdCP uses a task-first architecture where core advertising tasks can be accessed through multiple underlying protocols (MCP, A2A, or future protocols from AAIF).

Visit AdCP Documentation

Agentic Commerce Protocol (ACP)

Maintained by OpenAI and Stripe

ACP is an open standard for AI-powered commerce. It enables buyer agents, merchant systems, and payment providers to coordinate checkout flows, share credentials, and complete purchases in a secure, auditable way.

Key capabilities:

  • Product discovery and catalog navigation
  • Cart management and checkout coordination
  • Payment credential handling
  • Order fulfillment and tracking

ACP is open source under Apache 2.0 license, allowing any business to implement the specification to transact with any AI agent or payment processor.

Visit agenticcommerce.dev

Future Vertical Protocols

As the agentic ecosystem matures, additional vertical protocols will likely emerge for other domains such as service management, healthcare coordination, financial transactions, and more. Each will address the specific requirements and regulations of its domain while building on the shared horizontal infrastructure.

Traditional Standards

Agentic protocols don't replace existing industry infrastructure - they build on top of it. Traditional standards continue to provide essential capabilities that the agentic layer leverages.

IAB Tech Lab

Ad Tech Industry Standards

The Interactive Advertising Bureau (IAB) Tech Lab develops and maintains technical standards for the digital advertising industry. These standards form the underlying infrastructure that agentic advertising systems interact with.

Key standards:

  • OpenRTB - Real-time bidding protocol for programmatic advertising
  • VAST - Video ad serving template for video advertising
  • Open Measurement - Viewability and verification measurement
  • ads.txt / sellers.json - Supply chain transparency

When an AI agent executes a media buy through AdCP, the underlying transaction often flows through OpenRTB-compatible systems. Agentic protocols provide the AI-native interface; traditional protocols handle the execution.

Visit IAB Tech Lab

Complementary, Not Competing

It's important to understand that agentic protocols and traditional standards serve different purposes:

The agentic layer makes it possible for AI assistants to interact with advertising systems using natural language and intelligent workflows. The traditional layer ensures those interactions execute reliably across the existing ecosystem.

Get Involved

The agentic protocol landscape is being built in the open. Whether you're interested in horizontal middleware, vertical protocols for advertising or commerce, or traditional ad tech standards, there are opportunities to participate and shape the future.