Handoffs

It gets a human. With the whole story.

When the conversation needs a person, Outlearn hands it over with the full transcript and an AI summary, into Intercom, Zendesk, LiveChat, Slack or an inbox. The customer never repeats themselves. You decide when, in a sentence.

What your agent receives

How the copilot helps them finish

Replay

Customer

This is the third time my order's come through wrong. I want to speak to someone about compensation.

AI · Decision

This needs a human agent

Handoff · Complete

• Conversation context transferred
• AI summary attached
• Agent notified instantly

It hands off rarely. We publish how rarely.

Hand off everything and you have built an expensive contact form. Never hand off and you have built a wall. These are production numbers with the denominator attached, from the AI support benchmark.

• 29,003 real customer conversations, Oct 2025 to Aug 2026
• 45.9% resolved by the agent on its own, no human involved
• 5.7% handed to a person, with the transcript and a summary
• 2.4% handed off in the last 90 days, as the agent learned the docs

How it works

Four seconds between “I want a human” and a human

1. You say when, in a sentence
   • No rules engine. Write the trigger the way you would brief a new hire, and the agent makes the call.
2. The whole story travels
   • The person taking over receives the full chat transcript and an AI-written summary on top, so they know what was tried before reading a line.
3. Your agent is pinged instantly
   • The conversation lands where your team already works. An open thread is reused, never duplicated.
4. The customer repeats nothing
   • On a live handoff they stay in the same chat. On a ticket or email they are told who follows up, and when.

The trigger

Written like a note to a new hire

Every handoff has trigger instructions, and they are plain English. Specific enough that a password reset never escalates. Broad enough that an upset customer never gets a third automated answer. Run several, and route billing to one team and technical escalations to another.

Handoff · Support team

Active
Trigger instructions
Hand off to the support team when the customer:

• asks to speak with a human, a real person, or someone on the team.
• is visibly frustrated, or has raised the same issue more than twice.
• brings a billing dispute, a refund request, or a cancellation.
• asks something clearly outside what the agent should answer.

The agent reads the conversation and decides. Nothing else to configure.

Where it lands

Into the tools your team already has open

A handoff that lands in a new inbox is just another inbox. The agent answers on every channel you run and hands off into the tools you already have. The plan comparison shows how many destinations each tier allows.

• Intercom
• Live chat·Copilot
• Zendesk Sunshine
• Live chat·Copilot
• Zendesk Ticketing
• Ticket
• LiveChat
• Live chat
• Freshdesk
• Ticket
• HubSpot
• Ticket
• Jira
• Issue
• Slack
• Channel
• Email
• Inbox

Copilot: after the handoff, Outlearn drafts the reply from your knowledge and the transcript; your agent edits or sends. See the LiveChat setup for a walkthrough.

The difference

Most bots treat escalation as someone else’s problem.

A typical chatbot:

• “Please email support@”, the polite dead end.

Outlearn:

• Hands off inside the same conversation, or opens the ticket itself.
• The customer tells the whole story again.
• Transcript and AI summary arrive before the agent says hello.
• Escalation is a rules tree someone has to maintain.
• A trigger written in plain English, edited like a note.
• Billed per resolution, so a handoff is a lost sale.
• Credits for the AI’s work. A handoff costs nothing extra.

Pricing, for the record: credits for the AI’s work, never per seat, never per resolution. Free is $0 with 100 credits, Starter $99/month with 2,000, Growth $199/month with 4,500, Scale $599/month with 15,000. 14-day trial, no card.

The questions support leads ask about handoff

What is a chatbot human handoff?

The moment an AI agent decides a conversation needs a person and transfers it, with everything that was discussed, to a human.

Outlearn does it with the full transcript and an AI-written summary attached, notifies the agent instantly, and keeps the customer in the same conversation on a live handoff so nobody repeats themselves.

How does Outlearn decide when to hand off?

You tell it, in plain language.

Each handoff has trigger instructions you write yourself: hand off when the customer asks for a person, when they are frustrated, when it is a refund or cancellation, or when the question is outside the agent's knowledge. There is no rules engine to maintain.

What does my agent receive?

The full transcript, an AI-written summary of what was asked and tried, and an instant notification.

On a ticket handoff that arrives as a support ticket with the summary; on a live handoff the agent picks up the same chat. If the customer already has an open conversation with your team, Outlearn reuses it rather than opening a duplicate.

Does the human agent see the chat transcript?

Yes, all of it.

The full chat transcript travels with the handoff, with the AI summary above it, so the agent reads what the customer asked and what the agent tried before saying hello. On a live handoff they pick up the same chat; on a ticket handoff it arrives inside the ticket, and on an email handoff inside the email itself.

Can it hand off to a human in real time, with the full conversation history?

Yes.

On a live handoff the human picks up the same chat the customer is already in, with the full transcript and an AI summary above it, the moment the trigger fires; the agent is notified instantly. Where your team works from a queue instead, the same transcript and summary arrive as a ticket or an email. Either way the customer stays in one conversation and repeats nothing.

How often does it actually hand off?

Rarely, and on purpose.

Across 29,003 conversations from October 2025 to August 2026, the agent resolved 45.9% on its own and handed 5.7% to a human. In the last 90 days it was 50.8% resolved and 2.4% handed off. The method is published on the AI support benchmark, so you can hold us to it.

Does a handoff cost extra?

No.

Outlearn is priced in credits for the AI's work, never per seat and never per resolution, so a conversation that goes to a person is not billed as a win first. Plans start free with 100 credits, and the pricing page shows how many handoff destinations each tier allows.

How long does it take to set up?

Minutes for the first destination.

Open Handoffs in the dashboard, connect Intercom, Zendesk, LiveChat or an inbox, write the trigger in a sentence or two, and test it by asking the agent for a human. The LiveChat guide walks through the five steps, and none of them needs an API key.