A2A (Agent-to-Agent)

What it means

Agents pass messages, goals, or structured handoffs to peers (planner → researcher → writer) often via protocols, queues, or orchestration layers like MCP or custom APIs.

Why designers should care

Users need a simplified timeline of which agent did what, not a raw mesh of agent chatter. Surface roles, handoffs, failures, and final ownership clearly.

Example

A product brief flow shows Planner → Research agent → Draft agent → Review agent as a stepper; users can inspect each handoff packet and reject before publish.

Common mistakes

• • Exposing full agent-to-agent logs that read as noise to non-technical users.
• • Circular agent loops with no user-visible stop or cost cap.
• • Ambiguous accountability when multiple agents touch the same artifact.

Related terms

• Agent: The model decides sequences of actions (search, draft, click, call APIs) within policies until the task completes or hits a human checkpoint.
• Orchestration: An orchestrator decides what runs next: call search, pass results to the LLM, invoke a tool, branch on validation, or pause for approval.
• MCP (Model Context Protocol): MCP servers expose capabilities (search files, query tickets, run commands) that compatible clients discover and call in a standardized way.
• Workflow: Defined stages (inputs → generate → review → export) often spanning people, systems, and multiple model calls with clear deliverables.
• Claw: The claw metaphor describes how autonomous layers extend from chat into systems of record. Users feel it as background work, spawned sub-agents, and actions outside the main conversation thread.
• Subagent: The parent agent delegates a scoped goal to a subagent with its own tools, context limit, and completion criteria, then merges results or waits for approval.

Explore more

• Agentic UX framework