What it is
An Agent Handoff Protocol defines what information is passed, in what format, and under what conditions when an AI agent reaches the limit of its scope and transfers a task. This includes session context, intent history, collected data, and confidence scores. Without a formal protocol, handoffs between agents — or from agent to human — result in lost context and frustrated users.
Why it matters
Seamless handoffs are the difference between an agentic workflow that feels intelligent and one that feels broken. For customer-facing deployments, a poor handoff is indistinguishable from a system failure. Standardized protocols also make multi-agent systems auditable and easier to debug.
Key components
- context bundle
- trigger conditions
- receiving agent registration
- escalation path
- handoff logging
How it connects
In Agentforce, handoff behavior is configured through Topics and Actions, with escalation paths defined per topic. The Trust Layer logs handoff events for compliance and observability.
Good to know
Always include a human-readable summary in the handoff payload — it ensures human agents can take over without reading a full transcript.
Related terms
Human-in-the-Loop (HITL)
A design pattern where AI systems require human approval or intervention at critical decision points before taking action.
Multi-Agent System
An architecture where multiple specialized AI agents collaborate to handle complex workflows — each agent owns a domain and they hand off work to each other.
Topics and Actions
The building blocks of Agentforce agents — topics define what an agent can discuss, actions define what it can do.
Agent Observability
The practice of inspecting, debugging, and understanding AI agent behavior at runtime by consuming agent telemetry — traces, metrics, logs, and events — through dashboards, alerts, and evaluation tools.
Agent Orchestration
The coordination and management of multiple AI agents working together to accomplish complex workflows that no single agent could handle alone.