What is chatbot fallback rate?

Fallback rate, explained

Every chatbot needs a fallback: what it does when it cannot match a question to an answer. For rule-based bots, that is when no intent or keyword matches. For AI agents, it is when retrieval finds no relevant source or the system's confidence is below a threshold.

Fallback rate counts how often that happens. A high rate usually means the knowledge base has gaps, the bot is deployed on pages where visitors ask questions it was not prepared for, or the matching threshold is set too strictly.

A low fallback rate is not automatically good. An AI system that never falls back may be answering questions it should not, filling gaps with invented details. The goal is a fallback rate that reflects genuine gaps, paired with a low rate of unsupported answers.

The most useful thing about fallbacks is the list behind the number. Grouping fallback messages by topic shows exactly which content to write or which questions should go straight to a person.

How to measure it

State whether you count per message or per conversation; the two give very different numbers. Review fallbacks alongside a sample of confident answers.

Common mistake

How SuperBot relates

SuperBot analytics show intents, missed answers and knowledge gaps, and the chat hands off to a human when confidence is low, so fallbacks become a to-do list rather than a dead end.

• Knowledge gaps: the AI support metric that matters
• AI chatbot evaluation checklist

Frequently asked questions

Related terms

• Knowledge gapsKnowledge gaps are the customer questions that a support team's content, and therefore its AI agent, cannot answer. They show up as fallbacks, low-confidence answers, escalations and repeated rephrasing. Tracking them tells you exactly which help articles to write or fix next, making them one of the most actionable metrics in AI support.
• Hallucination (AI)An AI hallucination is a response from a language model that sounds confident and fluent but is false, unsupported by its sources, or invented outright. In customer support, typical hallucinations are made-up policies, prices, delivery dates or product features, and they are dangerous because customers act on them and businesses may have to honour them.
• Human handoff (escalation)Human handoff is the point where an automated system such as a chatbot passes a conversation to a person, ideally with the full conversation history and context attached. Good handoff rules trigger on low confidence, sensitive topics, explicit requests for a human or signs of frustration, and tell the customer what happens next and how long it will take.
• Resolution rateResolution rate is the percentage of customer conversations or tickets that end with the customer's issue solved. For AI support, it usually means conversations the bot resolved without a human. Because vendors define “resolved” differently, the number is only meaningful when you know the definition behind it and check it against repeat contacts and satisfaction.
• Intent detectionIntent detection is identifying what a customer is trying to achieve from their message, such as tracking an order, requesting a refund or asking about pricing, and labelling it as a category. Chatbots use it to choose a response or route a conversation, and support teams use intent reports to see what customers ask about most.