Transcribe Meeting and Send the Notes to Slack

When Recall completes transcribing a meeting, this trigger activates the workflow and captures {{recall.transcript}} which contains the full conversation with timestamps and speaker identification in the format "[hh:mm:ss] Speaker Name: text". The workflow runs for every meeting transcribed by Recall.

Setup requirement: You must have Recall configured to record and transcribe meetings.

This step sends the transcript to AI with specific instructions for analysis. The AI prompt directs it to:

1. Create a concise 2-3 paragraph summary of the conversation
2. Extract a clear list of any action items, decisions, or follow-ups mentioned
3. State "No action items found" if none exist

The AI receives {{recall.transcript}} (the full meeting transcript with timestamps and speakers) and processes it to understand conversation flow, identify commitments and decisions, and distinguish important points from casual discussion. The structured output is returned as {{ai.response}} with formatted sections for summary and action items.

AI capabilities: The AI understands meeting context like who committed to tasks, deadlines mentioned, decisions made, and can distinguish between actual action items versus casual mentions. It maintains speaker attribution for accountability.

This step posts the AI's analysis to Slack.

Configuration: You must select your Slack channel during setup—first invite the MESA Slack app to your desired channel by typing @MESA and clicking Invite (required for private channels), then select that channel in the configuration dropdown. The message contains {{ai.response}} (the formatted meeting notes with summary and action items). The markdown_text parameter enables formatting like bullet points, bold text, and sections that make notes scannable.

Message format: The AI returns structured notes with clear sections like "Summary:" and "Action Items:" that display as formatted Slack messages with proper spacing and bullet points.