The promise of AI support is that it handles the routine so humans can focus on the complex. That framing only works if there's actually a path from AI to human — a clear, working handoff when the agent reaches its limits.

Without one, users hit a wall. The agent can't help, there's no obvious next step, and they leave frustrated.

What handoff looks like in Companin

You configure what the path looks like. Options include:

Email form. The user fills out a short form — name, email, description — and your support team receives it as a ticket. Works with any email-based support system. No integration required.

Custom URL redirect. The agent presents a button that links to your support portal, Calendly, or any other destination.

Custom message with contact details. The agent shows a message you write — "Our team is available weekdays 9–5 at support@yourcompany.com" — and ends the conversation cleanly.

What triggers a handoff

Automatic triggers. When the agent's confidence falls below your threshold and no useful answer is available, it hands off instead of guessing.

Phrase triggers. Specific phrases you define always route to handoff, regardless of confidence. "I want to talk to a human", "I need a refund", "cancel my account." These are configured from the AI Behavior Editor. No code required.

Why the fallback experience matters

Users remember the edge cases disproportionately — the moment things went wrong and what happened next. If the edge case experience is "the bot couldn't help and then nothing," that's the impression users leave with. If it's "the bot couldn't help but immediately connected me to someone who could," that's a very different impression.