RecipeAgents / Guides
Agent recipe guideAI meeting notes to tasks automation
Meeting note tools already write summaries. What they don't do on their own is put action items where your team works. This recipe connects the transcript to your task tool so decisions and to-dos don't get lost in a recap email.
It uses Fathom or Otter for transcripts, Zapier for glue, and Notion or Asana for tasks.
The prompt
From this transcript extract: decisions, action items (owner, task, due date if stated), open questions. Return JSON. Transcript: {{transcript}}Step by step
- Trigger: transcript ready in Fathom or Otter.
- Send the transcript to an LLM step with the extraction prompt.
- For each action item, create a task in Asana or Notion with the owner and due date if one was stated.
- Post a short summary (decisions and open questions) to the team's Slack channel.
'If stated' is the key phrase
The prompt only asks for due dates that were said in the meeting. Without that, models tend to invent plausible deadlines. Leaving the date empty is better than a wrong one; the owner can set it.
Mapping owners
Transcripts use first names. Keep a small lookup table (name to Asana or Notion user) in the automation so tasks are assigned to the right person. Unmatched names can go to the meeting organizer.
Mistakes to avoid
Creating tasks for every 'we should' comment. Ask the model to include only items someone agreed to do.
Recording meetings without telling attendees; follow your local consent rules.
Posting full transcripts to busy channels. Post the summary and link the notes.
Build it in Make, Zapier or n8n
The recipe is tool-agnostic. Every step maps to a standard building block: a trigger from the app where the data starts, an AI step that sends text to a model with the prompt, a parser that reads the JSON or text that comes back, and actions in the destination apps. Zapier is the quickest to get running for linear flows. Make makes branching and loops easier to see. n8n can be self-hosted if you want to keep data on your own infrastructure. Pick the one your team already pays for.
Test it before you trust it
Run the recipe on ten to twenty real examples before you switch it on for everything. Put the inputs and outputs side by side in a sheet and mark each one right or wrong. Where it's wrong, add one specific line to the prompt that addresses the pattern, then run the same examples again. Keep a human approval step on any action that is visible to customers or touches money until the error rate is one you'd accept from a new hire.
Watch running costs as well. Each AI step is billed by the model provider and each task or operation by the automation tool. Filtering inputs before the AI step, for example skipping newsletters or tiny files, is the easiest way to keep both down. Check current pricing on each vendor's site, since it changes often.
Before you use it
Every recipe on RecipeAgents follows the same shape: a trigger, the tools, the steps, and the exact prompt. Start with the human-approval version. Once you have watched a few dozen runs and trust the output, decide which steps can run on their own. Model and tool names change often, so check current pricing and features in each tool before you build.
Frequently asked questions
- Which meeting note tool should I use?
- See our meeting notes ranking on TaskRank. Any tool that exposes a 'transcript ready' trigger works here.
- Notion or Asana?
- Use whichever your team already lives in. The recipe is the same.
- Can it run without Zapier?
- Make or n8n can do the same steps if your note tool connects to them.