What is human handoff?

Human handoff (escalation), explained

Automation handles the repeatable questions; people handle judgement, exceptions and emotion. Handoff is the bridge between the two, and it is where many chatbot experiences fail.

Common triggers include the customer asking for a person, the bot not finding a confident answer, sensitive topics such as billing disputes, complaints or safety issues, repeated rephrasing that suggests the bot is not helping, and high-value moments such as a large sales enquiry.

What gets passed matters as much as when. The receiving teammate should see the full transcript, what the bot already answered, any details collected such as order number or email, and why the handoff happened. Without that, the customer repeats everything and the automation has made the experience worse.

Tell the customer the truth about timing. If nobody is online, say so, collect contact details and set an expectation, rather than leaving the chat spinning.

Finally, review handoffs as a set. The reasons conversations were escalated show which content is missing, which rules fire too often and which topics should skip the bot entirely.

How to measure it

Also review a sample of non-escalated conversations for cases that should have been handed off.

Common mistake

How SuperBot relates

SuperBot hands chats off to your team in one shared inbox with the full conversation history attached. When confidence is low, the chat waits for a human, and escalation rules are configurable.

• Human handoff rules that keep an AI agent from annoying customers
• AI chatbot evaluation checklist
• SuperBot features

Frequently asked questions

Related terms

• Shared inboxA shared inbox is a single queue where a whole team can see, assign and reply to customer conversations, instead of messages sitting in one person's email. It shows who is handling what, keeps the conversation history in one place, and prevents both duplicate replies and conversations that nobody answers.
• Fallback rateFallback rate is the percentage of chatbot messages or conversations where the bot could not produce a confident answer and fell back to a default response, such as “I'm not sure” or an offer to contact a person. It measures how often customers ask things the bot's knowledge or rules do not cover.
• Live chatLive chat is a support channel where customers exchange real-time text messages with a business, usually through a widget on its website or app. Traditionally the replies came from human agents; many live chat tools now put an AI agent first and bring in a person when needed, so “live” increasingly means immediate rather than necessarily human.
• Sentiment analysisSentiment analysis is the automated detection of the emotional tone in text, typically classified as positive, neutral or negative. In customer support, it is applied to messages and conversations to spot frustrated customers, track how people feel about topics over time, and flag conversations for a human to review, without waiting for a survey response.
• First contact resolution (FCR)First contact resolution (FCR) is the percentage of customer issues fully resolved in the first interaction, with no follow-up, transfer or repeat contact needed. It is one of the most customer-centred support metrics because having to contact a business again about the same problem is a common source of frustration.