We have customers that answer no multiple times to Fins questions of ‘Did that answer your question’ or similar. They will say ‘No’ and rephrase their question but Fin still can’t answer it correctly. Does anyone have any recommendations on how I could get it to ask if they want to speak to a human after the 2nd time of saying ‘No it hasn’t helped’
Answered
How to deal with customers saying 'No' to Fins question of whether it answered their question
+1Best answer by Christopher Boerger
Hey Peter! Good news — there are actually a few complementary ways to handle this, and the docs have a great example for your exact use case.
Option 1 — Escalation Guidance (natural language)
Go to Train > Escalations > + New under Escalation Guidance and write something like:
"If a customer says 'No' or 'That didn't help' twice, offer to connect them to a human agent."
The docs even list "Offer to escalate the customer to a human agent if they ask the same question twice" as a built-in example. You can also phrase it as "escalate immediately" instead of "offer to escalate" if you want a hard handoff without the customer needing to confirm.
Option 2 — Escalation Rules (data-driven)
If you have structured data available (e.g. a Sentiment attribute, VIP flag, or Topic), you can use Escalation Rules instead — these trigger based on detected attributes rather than natural language. For the "No" use case, Guidance is likely the better fit, but Rules are worth knowing about.
Option 3 — The "question mark trick"
Worth knowing: Fin now generates its feedback question dynamically (e.g. "Was that helpful?" or "Did that answer your question?"). If you end a Guidance rule with a question mark, Fin will ask that question instead of its auto-generated one.
For example:
"At the end of every direct answer, ask 'Would you rather talk to a human?'"
This gives you more control over how the resolution check is phrased, which can reduce ambiguous "No" responses in the first place.
A couple of things to keep in mind:
- By default, Fin already detects a loop after 3 rounds of a customer repeating themselves — but Escalation Guidance lets you trigger escalation earlier (e.g. after the 2nd "No"), and it takes precedence over that default behavior.
- Be careful about making escalation offers too broad — the docs flag that generic guidance like "escalate if the customer is upset" can sharply increase your escalation rate and drop Fin's resolution rate.
References:
+3Hey Peter! Good news — there are actually a few complementary ways to handle this, and the docs have a great example for your exact use case.
Option 1 — Escalation Guidance (natural language)
Go to Train > Escalations > + New under Escalation Guidance and write something like:
"If a customer says 'No' or 'That didn't help' twice, offer to connect them to a human agent."
The docs even list "Offer to escalate the customer to a human agent if they ask the same question twice" as a built-in example. You can also phrase it as "escalate immediately" instead of "offer to escalate" if you want a hard handoff without the customer needing to confirm.
Option 2 — Escalation Rules (data-driven)
If you have structured data available (e.g. a Sentiment attribute, VIP flag, or Topic), you can use Escalation Rules instead — these trigger based on detected attributes rather than natural language. For the "No" use case, Guidance is likely the better fit, but Rules are worth knowing about.
Option 3 — The "question mark trick"
Worth knowing: Fin now generates its feedback question dynamically (e.g. "Was that helpful?" or "Did that answer your question?"). If you end a Guidance rule with a question mark, Fin will ask that question instead of its auto-generated one.
For example:
"At the end of every direct answer, ask 'Would you rather talk to a human?'"
This gives you more control over how the resolution check is phrased, which can reduce ambiguous "No" responses in the first place.
A couple of things to keep in mind:
- By default, Fin already detects a loop after 3 rounds of a customer repeating themselves — but Escalation Guidance lets you trigger escalation earlier (e.g. after the 2nd "No"), and it takes precedence over that default behavior.
- Be careful about making escalation offers too broad — the docs flag that generic guidance like "escalate if the customer is upset" can sharply increase your escalation rate and drop Fin's resolution rate.
References:
+1Thanks
+3Hi Petter Cool happy to hear so ;)
Join the Intercom Community 🎉
Already have an account? Login
Login to the community
No account yet? Create an account
Intercom Customers and Employees
Log in with SSOor
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.