An AI Catch-Up Assistant to Tackle Slack Information Overload
A Slack Catch-Up Assistant that sorts the unread pile into actions, context and decisions
✳The problem they went after
Knowledge workers come back from a vacation, a long meeting or a focus block to a wall of unread Slack messages. The diagnosis here is careful: volume is not the problem, missing structure is. Nothing tells you which decisions were made, what you now owe someone, or which announcements genuinely mattered, so catching up costs time and still leaves people unsure they caught up.
✳What they actually did
The assistant generates a personalised dashboard when a user returns, split three ways: a prioritised action items list pulled from direct mentions, assigned tasks and open questions; a need-to-know summary of high-impact announcements and changes; and a decisions and timeline view in chronological order. Natural language understanding and intent detection separate requests from informational messages, and the spec draws a line at making decisions for the user. The target audience is product managers, engineers, designers and people managers on high-volume channels.
✳What you can take from it
Worth stealing
- Diagnosis separates message volume from missing structure
- Three views: action items, need-to-know, decisions timeline
- Intent detection splits requests from informational messages
- Stated limit: the AI does not decide for you
Read this if
Speccing an AI summarisation feature inside a communication tool
The transferable bitSeparating volume from structure changes the solution space completely, because a summariser compresses text while this sorts by what the reader has to do about it. The explicit limit on deciding for the user is a good habit in AI specs, where scope drifts toward autonomy by default.
✳The submission itself
Open the PDF →Reading on a phone? The embedded viewer is cramped — open the PDF in a new tab instead.
