1What it does
Drop a voice memo or note in one Slack channel. Make transcribes it with Whisper, classifies the intent and sends it to the right model: Claude for content, Perplexity for research, GPT for summaries. Results come back as a threaded Slack reply and get logged to Sheets.
2Who it is for
- Industry
- Creators
- Platform
- Make
- Category
- Operations
- Tools used
- MakeSlackOpenAI WhisperClaudePerplexityGoogle Sheets
3How it works
Every step, in order, is in Workflow at a glance, pinned on the right as you scroll.
- 01 Trigger
New message in a Slack channel (webhook).
- 02 Action
Router detects voice memo vs text.
- 03 AI / action
Download the file and transcribe with OpenAI Whisper.
- 04 AI / action
AI categorization: social media, email, research or brain dump.
- 05 AI / action
Route: Claude writes posts and emails, Perplexity runs research, GPT summarizes brain dumps.
- 06 Action
Log to Google Sheets with formatted timestamps.
- 07 Action
Reply in the Slack thread.
- Trigger
- New message in a Slack channel (webhook).
- AI step
- Classify this message into exactly one of: social, email, research, braindump. Return only the label.
4How it went
Built live on stream. Aaron switched from a Make Agent to a plain text categorization step for the classifier. No usage results reported.
Only what the source reports. The source did not report numbers.
5Worth knowing
- A plain classification step was simpler than a Make AI Agent for routing.
- Set variables before the router so every branch can read them.
6Recipe
Written by Builds That Work The source did not publish this. We wrote it as a starting point.
Classifier prompt: "Classify the message below into exactly one label: social, email, research, braindump. Return only the label, lowercase.\n\nMessage: {{transcript}}"7Source
Open the original YouTube by Aaron Delz (ClickLaunch) →
Source opened and read on 2026-10-11.