Can I format notes for different meeting types differently?

Yes, add a router module after the Slack trigger that checks for keywords like 'standup' or 'client call' and routes to different OpenAI modules with specialized prompts. Each route can have its own formatting template.

What happens if someone posts a really long meeting transcript?

OpenAI has a 4,096 token limit for GPT-3.5 and 8,192 for GPT-4. Add a text truncation module before the AI call to cut messages at 3,000 tokens (roughly 2,400 words) to avoid errors.

Can I prevent the bot from processing its own formatted responses?

The 'Watch Bot Messages' setting in the Slack module should be unchecked. If you're still getting loops, add a filter that excludes messages containing your formatted output structure like 'Action Items:'.

How do I handle private channels or DMs?

Make's bot needs to be invited to private channels manually. For DMs, use the 'Watch Direct Messages' trigger instead of 'Watch Messages' and configure it the same way.

Beginner~12 min setupAI & CommunicationVerified April 2026

How to Clean Meeting Notes with OpenAI and Slack with Make

Automatically transform raw meeting notes pasted in Slack into structured summaries with action items using OpenAI's GPT-4.

Steps and UI details are based on platform versions at time of writing — check each platform for the latest interface.

Make

Also on:Zapier N8n Power-automate Pipedream

Best for

Teams that regularly paste messy meeting notes into Slack and want them automatically cleaned up with action items extracted.

Not ideal for

Teams that already take structured notes during meetings or only have 1-2 meetings per month.

Sync type

real-time

Use case type

notification

Real-World Example

💡

A 12-person marketing agency uses this to clean up client meeting notes posted in their #client-updates channel. Account managers paste raw notes after calls, and the bot reformats them into summary, action items, and decisions within 30 seconds. Before automation, the team spent 15 minutes after each meeting manually formatting notes, and action items often got buried in rambling paragraphs.

What Will This Cost?

Drag the slider to your expected monthly volume.

Monthly workflow runs

/mo

505005K50K

Each platform counts differently — Zapier: 1 task per trigger. Make: 1 operation per module per record. n8n: 1 execution per run.

$30/mo if billed monthly

Pipedream

$29/mo

Basic plan

Prices shown for annual billing. Based on published pricing as of April 2026.

Estimated ROI

1000

min saved/mo

$583

labor value/mo

Free

no platform cost

Based on ~2 min manual effort per operation at $35/hr fully loaded labor cost.

Implementation

Skip the setup

Import this workflow directly into Make

Copy the pre-built Make blueprint and paste it straight into Make. All modules, filters, and field mappings are already configured — you just need to connect your accounts.

Before You Start

Make sure you have everything ready.

Slack workspace admin access to install Make app

OpenAI account with API access and available credits

Make account with available operations quota

Optional

Dedicated Slack channel for meeting notes processing

Team agreement on trigger phrase format

Field Mapping

Map these fields between your apps.

Field	API Name	Description
Required
Message Text	`text`	Raw meeting notes content from the Slack message to be processed by AI
Channel ID	`channel`	Slack channel identifier where the response should be posted
Thread Timestamp	`thread_ts`	Original message timestamp to create threaded reply
AI Response	`content`	Formatted meeting notes output from OpenAI GPT-4 model

2 optional fields▸ show

User ID	`user`	Slack user who posted the original meeting notes
Message Timestamp	`ts`	When the original meeting notes were posted

Step-by-Step Setup

Dashboard > + Create scenario > Slack > Watch Messages

Create new scenario

Start a new Make scenario that will listen for Slack messages and process them through OpenAI. This sets up the foundation for your meeting notes cleanup automation.

1Click 'Create a new scenario' from your Make dashboard
2Click the gray circle with a '+' to add your first module
3Search for 'Slack' in the app list
4Select 'Watch Messages' from the Slack trigger options

✓ What you should see: You should see a Slack module with 'Watch Messages' selected and a connection prompt.

Slack module > Connection > Add

Connect Slack workspace

Link your Slack workspace to Make so it can monitor specific channels for meeting notes. You'll need admin permissions to install the Make app.

1Click 'Add' next to the Connection field
2Click 'Authorize' to open Slack's OAuth flow
3Select your workspace from the dropdown
4Click 'Allow' to grant Make access to your Slack workspace

✓ What you should see: The connection field should show your workspace name with a green checkmark.

⚠

Common mistake — Make sure you have permission to install apps in your Slack workspace — regular members can't complete this step.

Make settings

Connection

Choose a connection…Add

click Add

OpenAI

Authorize Make

popup window

✓

Connected

green checkmark

Slack module > Channel selection

Configure channel monitoring

Set Make to watch a specific channel where people will paste raw meeting notes. This prevents the bot from processing every message in your workspace.

1In the 'Channel' dropdown, select your target channel (e.g. #meeting-notes)
2Leave 'Watch Bot Messages' unchecked to avoid loops
3Set 'Limit' to 1 to process one message at a time
4Click 'OK' to save the module configuration

✓ What you should see: The Slack module should show your selected channel name and be ready for testing.

⚠

Common mistake — Don't select a high-traffic channel like #general — you'll burn through operations processing irrelevant messages.

Between modules > Add filter

Add message filter

Create a filter that only processes messages containing meeting notes, not every chat message. This saves operations and prevents spam.

1Click the wrench icon between modules to add a filter
2Name the filter 'Meeting Notes Only'
3Set condition to 'Text contains' and enter 'meeting notes:' or your trigger phrase
4Click 'OK' to apply the filter

✓ What you should see: You should see a small funnel icon between the Slack module and the next step.

⚠

Common mistake — Pick a consistent trigger phrase your team will use — changing it later means updating the filter and retraining users.

Message template

📬 New entry: {{1.name}}
Email: {{1.email}}
Details: {{1.description}}

OpenAI

trigger

filter

Condition

matches criteria?

yes — passes through

no — skipped

Slack

notified

Add module > OpenAI > Create a Completion

Add OpenAI module

Connect OpenAI's GPT-4 model to process the raw meeting notes. This module will send the Slack message content to OpenAI for restructuring.

1Click the '+' icon after your filter to add a new module
2Search for and select 'OpenAI'
3Choose 'Create a Completion' from the available actions
4Select 'gpt-4' from the Model dropdown

✓ What you should see: The OpenAI module should appear with GPT-4 selected and connection fields visible.

⚠

Common mistake — GPT-4 costs more than GPT-3.5 but gives much better results for structured note formatting — don't cheap out here.

OpenAI module > Connection > Add

Connect OpenAI API

Link your OpenAI account using an API key. You'll need a paid OpenAI account since the free tier doesn't include API access.

1Click 'Add' next to the Connection field
2Go to platform.openai.com/api-keys in a new tab
3Click 'Create new secret key' and copy the key
4Paste the API key into Make's connection field and click 'Save'

✓ What you should see: The connection should show 'Verified' status with your OpenAI account email.

OpenAI module > Messages configuration

Configure AI prompt

Write the prompt that tells GPT-4 how to restructure meeting notes. This is the core logic that transforms messy notes into clean summaries.

1In the 'Messages' field, select 'User' as the role
2In the message content, write: 'Format these meeting notes into: 1) Summary (2-3 sentences), 2) Action Items (bulleted list with owners), 3) Decisions Made. Raw notes:'
3Click the data picker and select 'Text' from the Slack module
4Set Max Tokens to 500 and Temperature to 0.3

✓ What you should see: The prompt should show your formatting instructions followed by the Slack message content variable.

⚠

Common mistake — Keep Temperature low (0.3 or less) for consistent formatting — higher values make the output too creative and inconsistent.

Add module > Slack > Create a Message

Add Slack response module

Send the formatted notes back to Slack as a thread reply. This keeps the original message and cleaned version together for context.

1Click '+' after the OpenAI module to add another module
2Select 'Slack' again from the app list
3Choose 'Create a Message' action
4Use the same connection from step 2

✓ What you should see: A new Slack module should appear configured to send messages to your workspace.

Slack Create Message > Channel and text configuration

Configure thread reply

Set up the response to post as a thread reply to the original message. This keeps your channel clean and makes it easy to find both versions.

1Select the same channel as step 3
2In the 'Text' field, click the data picker and select 'Content' from the OpenAI module
3In 'Thread TS', select 'TS' from the original Slack message
4Check 'Reply Broadcast' to make the response visible in the main channel

✓ What you should see: The module should show your channel, the AI response content, and thread settings configured.

⚠

Common mistake — Don't forget Thread TS — without it, you'll spam the main channel instead of creating organized threads.

Scenario controls > Run once

Test the workflow

Run a test with real meeting notes to verify the entire flow works correctly. This catches configuration issues before going live.

1Click 'Run once' at the bottom of the scenario
2Go to your Slack channel and post 'meeting notes: [paste some sample notes]'
3Return to Make and watch the execution log
4Check Slack for the formatted response in a thread

✓ What you should see: You should see a successful execution in Make and a clean, formatted response thread in Slack.

⚠

Common mistake — Use realistic test data — simple notes like 'we talked about stuff' won't show if the AI formatting is working properly.

Make

▶ Run once

executed

✓

OpenAI

✓

Slack

🔔 notification

received

Scenario controls > Scheduling toggle

Enable auto-scheduling

Turn on automatic execution so the scenario runs continuously without manual intervention. This makes your meeting notes cleanup fully automated.

1Click the 'OFF' toggle in the bottom-left to turn it 'ON'
2Set the schedule to 'Immediately' for real-time processing
3Click 'OK' to confirm the scheduling settings
4Save the scenario with a descriptive name like 'Meeting Notes Cleanup'

✓ What you should see: The scenario status should show 'ON' and 'Immediately' with a green indicator.

⚠

Common mistake — Don't set a polling interval — 'Immediately' uses webhooks for instant processing when messages arrive.

Drop this into a Make custom function.

Copy this template{{replace(replace(1.content; "\n\n"; "\n"); "**"; "*")}}

▸ Show code

{{replace(replace(1.content; "\n\n"; "\n"); "**"; "*")}}

... expand to see full code

{{replace(replace(1.content; "\n\n"; "\n"); "**"; "*")}}

Scaling Beyond 100+ meeting notes/month+ Records

If your volume exceeds 100+ meeting notes/month records, apply these adjustments.

Batch processing setup

Switch from immediate triggers to 5-minute polling intervals. This groups multiple meeting notes into single executions and reduces operation costs by up to 40%.

Model optimization

Use GPT-3.5-turbo for simple formatting and reserve GPT-4 for complex meetings. Add a router module that checks message length — under 500 words goes to the cheaper model.

Going live

Production Checklist

Before you turn this on for real, confirm each item.

Set error handling to 'Break' on OpenAI module to avoid malformed responses

Applies when: Always

Add rate limit retry with 30-second delay for OpenAI API calls

Applies when: High volume

Test with messages over 2000 characters to verify token limits

Applies when: Always

Configure webhook timeout to 120 seconds for long AI processing

Applies when: Always

Set up Make notification emails for scenario failures

Applies when: Enterprise

Create backup scenario with GPT-3.5 if GPT-4 fails

Applies when: High volume

Troubleshooting

Common errors and how to fix them.

Frequently Asked Questions

Common questions about this workflow.

Analysis

VerdictWhy Make for this workflow

Use Make for this if you want real-time processing and complex AI prompt management. Make's visual editor makes it easy to modify the OpenAI prompt and add logic like different formatting for different meeting types. The built-in error handling catches API failures gracefully. Skip Make and use Zapier instead if your team needs this running in under 10 minutes — Zapier's OpenAI integration is more plug-and-play.

Cost

This workflow burns 3 operations per meeting note: one Slack trigger, one OpenAI API call, one Slack response. At 50 notes per month, that's 150 operations total. Make's free tier gives you 1,000 operations monthly, so you're covered. The $9/month Core plan bumps you to 10,000 operations. Zapier would cost $20/month for the same volume since their AI actions count as premium. N8n is free but requires hosting.

Tradeoffs

Zapier wins on setup speed — their OpenAI integration has pre-built prompts for meeting notes and connects in 3 clicks. N8n gives you more control over the AI parameters and lets you switch between different models mid-workflow. But Make hits the sweet spot for most teams. The visual builder makes prompt tweaking easy, and the pricing stays reasonable until you're processing hundreds of meetings monthly.

OpenAI's API sometimes returns malformed JSON when asked for structured output — your formatted notes might break Slack's markdown. Add a text formatter module after OpenAI to clean up the response before posting. The Slack Watch Messages trigger has a 2-minute delay on free Slack workspaces, so responses aren't truly instant. If your team posts multiple meeting notes quickly, Make might process them out of order — add a delay module to space out the API calls.

Ideas for what to build next

→
Add action item tracking — Create a follow-up scenario that extracts action items from formatted notes and creates tasks in Asana or Monday.com with due dates and assignees.
→
Meeting summary dashboard — Send all formatted meeting notes to a Google Sheet or Airtable base to create a searchable archive and monthly summary reports.
→
Calendar integration — Connect to Google Calendar or Outlook to automatically tag meeting notes with the actual meeting title and attendees for better organization.