The biggest problem with the current automated approaches until today is the lack of context they have - they’re purely action/time based, which in some cases adds additional frustration to the person receiving the notification.
“Ugh - I didn’t find the solution I was looking for, and now you’re asking me if I found a solution and to mark it”
I lean slightly towards the side that this does more harm than good.
All that said - a lot has changed in recent years, and I believe AI to be a strong tool for enriching these types of flows in a way that makes it more meaningful to the person receiving the notification.
Combing an AI agent with an MCP, you open a world of opportunities:
-
After X days of inactivity within a topic, the agent reads the topic, uses the MCP (which has context of documentation and the rest of a community) and identifies posts that could be the solution. If it finds a reply, it uses that to send the notification to the OP to say, “I found this reply in a topic you created that seems to solve it - is this accurate? Would you like to mark it as the solution?”
-
Similar to the above, providing Admins a dashboard/report/table of topics that the Agent understands to have a solution for Admins to review and mark as solved - with reasoning for why they think it’s solved
-
It would also help in the situations where people indicate that something is the solution - “That’s great, and solved what I was looking for” - but don’t mark it as the solution. It’s a strong signal for AI to know there is a solution/answer.
I don’t know what the full UI/UX would look like - but am a strong believer that AI can and should play a role to provide better indications of what solutions are for both Admins and OPs