Front Macro-Gap Detector: Drafted Macros into a Coda Approval Queue

Goes a step past detection and actually writes the macro. For each cluster of repeated hand-typed answers, an LLM composes a clean, reusable canned response with a suggested title and tags, then logs it to a Coda table as a pending proposal your team can edit and approve in one place.

Use it when you want finished macro drafts waiting for review rather than just a list of topics. Ideal for teams that manage their knowledge and saved-reply backlog inside Coda and prefer a structured approval queue over chat threads.

1. 1A schedule triggers the run on your chosen cadence.
2. 2It collects recent manual Front replies and excludes ones sent from existing macros.
3. 3An OpenAI step groups them and, for each cluster, drafts a polished macro body plus a title and tags.
4. 4A logic step skips any draft whose meaning closely matches an existing macro to avoid duplicates.
5. 5Each surviving draft is written to a Coda approval table as a new row with status "Pending review," ready for a human to approve or reject.