What are the best practices for making Fin AI proactive? | Community
Skip to main content
Question

What are the best practices for making Fin AI proactive?

  • August 10, 2026
  • 0 replies
  • 33 views

Forum|alt.badge.img+1

Anyone else making Fin proactive? How are you triggering it — quick replies or the banner with the three options?

We've connected live user activity from Amplitude into Fin soand are using it to make Fin proactive based on real-time behavior. Trying to figure out the best way to actually surface that.

So far I've tried:

  • Quick replies — feels like the more unstable/experimental route (schema reference), and the options end up as pretty terse one-liners
  • The banner (three-option version) — more stable, but too small to show the full message before engaging

We want to find a solution that lets us show a bit more text in the proactive message — something in between: more room than the banner gives, but without shrinking down to quick-reply one-liner options.

Curious how others are approaching this — which method have you had the most luck with, and any gotchas I should know about before going further down one path?