AI Agent Handoff — How It Works

Agent handoff is how a conversation moves between specialized agents — or to a human — without losing context, so the visitor never repeats themselves and every answer stays grounded and cited. It is what makes a multi-agent platform feel like one coherent assistant instead of a wall of bots.

Why handoff matters

A single conversation often crosses boundaries: a support question becomes a sales question, a technical detail surfaces, billing comes up. Without handoff, the visitor either gets a vague generalist answer or is told to "contact a different team." With handoff, the support agent recognizes the shift and passes the thread — and its full context — to the sales or technical agent best suited to answer, mid-conversation and invisibly to the visitor.

How it works

Each agent has a defined scope and a set of tools. When the conversation moves outside an agent's scope, the orchestrator routes it to a better-fit agent, carrying the message history forward so nothing is lost. The visitor sees a continuous chat; behind the scenes, the right specialist is now answering. Every answer still cites its source, and the analytics dashboard records where handoffs happen so you can see which boundaries your visitors cross most.

Handing off to a human, with approval

Some conversations should reach a person — a hot lead, an edge case, a write that needs judgment. The agent captures a lead or escalates, and any action that changes data still pauses for human write-confirm. Handoff is one capability of the broader multi-agent platform, alongside grounding, citations, and MCP tools.

Getting started

Enable more than one agent, define each one's scope, and let the platform route between them. Try it live in the playground, then start free — 100 messages a month on the free tier, with room to grow on pricing.