What changed when Pupil Progress pushed Intercom’s Fin to the front of their support flow and rewrote their help centre for humans and AI? In this episode of Support Stack, Conor Pendergrast (CustomerSuccess.cx) interviews Tobi Davis (Head of Customer Experience, Pupil Progress) about how Tobi’s team lifted Fin’s resolution rate from ~55% to ~75% in about two months, and why it now holds steady at 75–80% weekly. You’ll learn:
-
Why going AI-first with a clear “talk to a human” option lets users self-select the path they want
-
How to rewrite your help centre for two audiences: people and Fin
-
The role of audiences and attributes so answers change by role, subject or context
-
When to use custom answers, snippets and narrow articles for precise guidance
-
Why they focused on content and data before actions or procedures
-
How tools like CX Score and Topic Explorer guide the next set of improvements
Who this helps:
-
Small support teams trying to scale quality without growing headcount
-
Intercom users who want practical steps to improve Fin performance
-
Leaders who prefer clear, repeatable levers rather than one-off tricks
If you want an actionable, real-world recipe for better AI resolution rates, this one is for you.
