What should a morning AI brief contain besides a summary of messages?
Teresa Carlson described a morning workflow that helps her review missed Slack conversations, important responses and things left unfinished the day before, and she said it needs corrections. What carries over is the shape of the work: a brief built to find decisions in a pile of messages, with the original message beside every item.
For a first version, define three sections. One is information to review, one is replies to consider, and one is unfinished work to check. Keep proposed actions separate from facts.
"The project lead requested a decision by noon" is a fact to verify in the source. "Reply before the team meeting" is a proposed next step. You may agree with it, but the brief shouldn't blur the line. Keep the first build to sources you're allowed to use and actions you'll review. A brief shouldn't send responses or change commitments just because it can spot them.
Then test it against your own review of the morning's work. See what it missed, what it ranked too highly and what it repeated after you'd handled it. Those corrections make a better assignment for the next run.