Hi all!
I recently read Richard Millington’s piece on “community hygiene” and how the long-term value of mature communities increasingly depends on AI tools surfacing the right answers, not just popular ones: Bad Community Hygiene Is Starting To Cause Big Problems.
One part really stuck with me: the idea that AI needs clear signals of truth/authority, structured information, and freshness signals to avoid confidently surfacing outdated or conflicting guidance
That sent me down a rabbit hole on our “Best Answer” behavior, and I’d love advice from folks here.
What I’m trying to understand (AI surfacing + “Best Answer”)
When an AI assistant (ChatGPT, Google’s AI experiences, in-product community AI, etc.) pulls from community threads, how much does “Best Answer / Accepted Solution” influence what gets surfaced?
A bit of context on us: we’re a B2B SaaS community made up of customers, with a Support team helping moderate and answer product questions.
Today, roughly half of our Best Answers are coming from peers. We’ve often picked the peer reply as Best Answer to reward awesome community behavior even when Support later posts a more complete “wrap-up” / definitive resolution, with helpful links, etc.
We’ve traditionally marked the peer reply (not the Support one) as Best Answer to encourage the helpful user-helping-user behavior… (they would get points, badges, “best answer” count going up...
but I’m wondering if that’s accidentally sending the wrong “this is the source of truth” signal for AI discovery.
Note: we don’t use the Knowledge Base in our Community. We use a separate Knowledge Base Tool for a “Single Source of Trouth” then the Community is there for troubleshooting, questions, best practices etc.
Also we have our own AI Support Assistant built-in to our tool. That one only pulls from our Help Center which is always up to date.
