What is a chatbot fallback response?
Quick answer
A chatbot fallback response is the set reply an AI chat assistant gives when it cannot find the answer in the material you gave it. Instead of guessing, it says it is not sure and tells the customer how to reach a person, such as an email address or phone number.
Last updated
Take the free module →Module 33 of the free AI Apps Worth Knowing course, about 14 minutes. No account needed.
Updated · By Robert Breen
Why it matters for a small business
Without a fallback, an AI chat tends to fill gaps with something that sounds right. For an office that is a real risk: a made-up price, a refund policy that does not exist or a promised deadline can turn into a complaint or a dispute. The fallback is the cheapest protection you can add, because it is one sentence in the instructions.
It also keeps the customer moving. A visitor who hears "I'm not sure, here is who can help" still gets somewhere. The fallback is the simplest form of human handoff, and every time it fires, it shows you a question your FAQ does not cover yet. Read those chats and add the missing answers.
In a real lesson: Chatbase: Answer Customer FAQs with an AI Agent
In the Chatbase FAQ agent lesson, Juniper Lane Bookkeeping (made up) writes its fallback into the Guardrails box under Build, Instructions. The rule says: if the answer is not in the training data, say "I'm not sure about that one. Please email hello@juniperlane.example or call (555) 014-2290 and a person will help." Another Guardrails line tells it never to guess prices, dates or policies.
The lesson keeps Temperature low and then tests the fallback on purpose. In the Playground you ask "Do you work with nonprofits?", a question the FAQ does not cover. The agent should say it is not sure and point to the email and phone number. If it makes something up, you tighten the Guardrails and test again. Making something up is a hallucination; the fallback is how you stop it from reaching a customer, alongside wider guardrails.
Try this lesson free or read the step-by-step guide.
Common confusions
Chatbot fallback response vs fallback output
The n8n fallback output is an extra exit on a Switch node for items that match no rule. A chatbot fallback response is words the customer reads in a chat.
Fallback response vs hallucination
A hallucination is the AI stating something false as fact. A fallback response is the instruction that tells it to admit a gap instead.
Tips
- Put real contact details in the fallback, not "contact us".
- Test it with two or three questions you know the FAQ does not answer.
- Keep the wording friendly and short; the customer should not feel blocked.
- Review chats where the fallback fired, and add the answers you can.
Related terms
More AI agents terms
Where you use it: free lessons
- Chatbase: Answer Customer FAQs with an AI Agent (Chatbase, 14 min)
- tawk.to: Answer Website Questions After Hours (tawk.to, 14 min)
AI tools that use it
Frequently asked questions
- Where do I write a chatbot's fallback response?
- In the bot's instructions. In Chatbase that is the Guardrails box under Build, Instructions; other tools have a system prompt or a "when the agent can't help" setting.
- Will a fallback stop every made-up answer?
- No, it makes them much less likely. Keep testing questions the bot should not know and read the chat logs regularly.
- Should the fallback ask for the customer's details?
- It can, but only basic contact details. Never ask for bank statements, passwords or Social Security numbers in a chat.