Hi all — I'm doing some research on how teams measure AI resolution quality, and there's one thing I can't find anyone talking about.
My understanding of the billing mechanic: a conversation counts as resolved either when the customer confirms it helped, or when they stop responding without asking for a human. The second case is the interesting one. Intercom does deduct the resolution if the customer reopens *that conversation* later, even across billing periods, which seems like a genuinely fair correction.
But it only catches people who come back to the same thread. If someone reads Fin's answer, decides it didn't help, and emails support@ three days later — or calls, or starts a fresh chat — that's a new conversation. Nothing connects it to the original, and from what I can tell there's no way to see that pattern from inside Intercom, because Intercom can only see the Intercom channel.
So my question: **has anyone here actually joined their email or phone support data against Fin's resolution log to see how often this happens?**
If you have, I'd love to know roughly what you found — even a rough sense of whether it's 2% or 20% would be useful. If nobody's measured it, that's a useful answer too.
Full disclosure: I'm exploring building something in this space, so I have a horse in the race. I'm not selling anything here and I'm not going to DM anyone — I genuinely just want to know if this is a real problem or something I've invented in my head.
