The Agent2Agent Protocol provides a common way for one agent to interact with another even when they were built with different frameworks or belong to different organizations. It supports capability discovery and structured task communication without requiring either agent to expose its internal implementation.
In agentic commerce, a customer agent can use A2A to coordinate with a merchant agent. The protocol carries agent-to-agent interaction, while commerce protocols define shopping operations and payment protocols constrain the transfer of money.
ELI5
The Agent2Agent Protocol gives independently built AI agents a common way to discover what each can do, exchange messages and coordinate work. They can cooperate without needing to reveal all of their internal code.
For example, a travel agent could ask a booking agent which reservation tasks it supports and then send a structured request. The protocol helps the agents communicate, while separate payment rules still control whether and how money may be spent.
What does A2A stand for?
A2A refers to Agent2Agent, also written Agent-to-Agent.
How is A2A different from MCP?
A2A focuses on communication and task coordination between agents, while MCP standardizes how AI applications connect to tools, resources and prompts.
