A2A
Agent-to-Agent Protocol
In one line
A2A (Agent-to-Agent Protocol) is the agent-to-agent standard Google launched and donated to the Linux Foundation in mid-2025, letting agents from different vendors delegate work to each other and exchange results.
Going deeper
A2A is the agent-to-agent protocol Google introduced at Cloud Next in April 2025 (and donated to the Linux Foundation that June). Where MCP targets the LLM-to-tool link, A2A targets the agent-to-agent link — one agent delegating work to another vendor's agent and consuming the result.
For marketers it sketches a near-future shape: a user's shopping agent talking directly to a brand's customer-service agent to check stock or negotiate price. That handoff is the canonical A2A example in industry talks.
MCP, A2A and UCP all landed in a short window, which makes the standards landscape feel crowded. The current read is that they live at different layers rather than competing head-on, but it will probably take through late 2026 to see which survive.
Sources
Related terms
MCP
MCP (Model Context Protocol) is the open standard introduced by Anthropic for connecting LLMs to external tools and data sources in a consistent way.
AI AgentUCP
UCP (Universal Commerce Protocol) is the AI-agent payment and checkout protocol Google introduced at NRF 2026, aimed at standardising how agents buy products on a user's behalf.
AI AgentMulti-Agent System
A multi-agent system is several AI agents with different roles cooperating on the same task — used when one agent alone is not enough to solve the problem.
AI AgentOrchestrator Agent
An orchestrator agent is the top-level agent that schedules sub-agents, assigns roles and enforces stop conditions — the control tower of a multi-agent system.
AI AgentAgentic Commerce
Agentic commerce is the emerging model in which AI agents handle the entire shopping flow — search, compare, decide, pay — on behalf of the user.