Build a Meeting Bot That Writes Your Notes For You Using AI

LinkBlogs

How to Build an Automated Meeting Summary Bot Using Make.com and OpenAI (No Code Required)

Step by step guide to automating meeting summaries using Make.com, OpenAI Whisper for transcription, and GPT-4o for structured summaries. Includes action item extraction, email delivery, and Notion integration.

Hey everyone! 👋

Okay so this one is for anyone who has ever sat through a one hour meeting, taken zero notes, and then spent another 30 minutes trying to remember what was actually decided.

That was literally me every week until I built this. And the best part? I am not a developer. I did not write a single line of code. I used Make.com (formerly Integromat) and OpenAI to build a bot that automatically transcribes my meeting recordings, summarises them, pulls out action items, and sends everything to my email and Notion workspace.

Here is the complete tutorial from start to finish. Follow along and you will have this running by the end of the post.

What This Bot Actually Does 🎯

Before getting into the steps, let me show you what the finished automation does so you know exactly what you are building:

You finish a meeting and upload the recording (audio or video file) to a Google Drive folder
Make.com detects the new file automatically
The audio gets sent to OpenAI's Whisper API for transcription
The transcript gets sent to GPT-4o with a custom prompt
GPT-4o produces a structured summary including key decisions, action items, and next steps
The summary gets sent to your email automatically
A formatted version gets saved to a Notion page automatically

All of this happens within about two to three minutes of uploading your recording. Zero manual work after the initial setup.

What You Need Before Starting 🛠️

Here is everything you need to have ready:

A Make.com account (free tier works for testing, Starter plan at around $9/month for regular use)
An OpenAI account with API access (pay-as-you-go, very affordable for audio transcription)
A Google account with Google Drive
A Notion account (free tier is fine)
Your meeting recording in MP3, MP4, M4A, or WAV format

That is it. No coding. No servers. No technical background needed.

Step 1: Set Up Your Google Drive Trigger Folder 📁

The automation starts when you upload a file to a specific Google Drive folder. So first we need to create that folder.

Go to your Google Drive and create a new folder. Name it something obvious like "Meeting Recordings" or "Bot Inbox." This is where you will drop your meeting recordings after every call.

Keep this folder link handy. You will need it in Step 3.

Step 2: Get Your OpenAI API Key 🔑

Go to platform.openai.com and sign in. Click on your profile in the top right and select "API Keys." Click "Create new secret key," give it a name like "Meeting Bot," and copy the key immediately because you will not be able to see it again after closing the window.

Paste it into a temporary note somewhere safe. You will need it in Step 4.

Cost note: OpenAI's Whisper transcription costs $0.006 per minute of audio. A one hour meeting costs roughly $0.36 to transcribe. GPT-4o summarisation adds another few cents depending on transcript length. Total cost per meeting: well under $1 in most cases.

Step 3: Create Your Make.com Scenario

Log into Make.com and click "Create a new scenario." You will see a blank canvas with a circle in the middle. This is where you build the automation visually by connecting modules together.

Adding Module 1: Google Drive Watch Files

Click the circle and search for "Google Drive." Select the "Watch Files" trigger. This module watches your folder and fires the automation whenever a new file appears.

Connect your Google account when prompted. Then configure:

Folder: select your "Meeting Recordings" folder
Watch: New files only
Maximum number of returned files: 1

Click OK to save this module.

Step 4: Add the OpenAI Transcription Module

Click the small circle on the right edge of your Google Drive module to add the next step. Search for "OpenAI" and select the "Create a Transcription" module.

Connect your OpenAI account by pasting the API key you copied in Step 2.

Configure the module:

File: select the file data from the Google Drive module (Make.com will show this as a variable you can click to insert)
Model: whisper-1
Language: en (or your language code)
Response format: text

Click OK. This module takes your audio file and converts it to text automatically.

Step 5: Add the GPT-4o Summary Module

Click the circle on the right edge of the transcription module. Search for "OpenAI" again and select "Create a Completion."

This is the most important configuration step. Here is the exact prompt setup that produces clean, structured meeting summaries:

Model: gpt-4o

System message:
"You are a professional meeting summarizer. Your job is to analyze meeting transcripts and produce clear, structured summaries that busy professionals can read in under two minutes."

User message (paste this exactly):

"Here is the transcript of a meeting: [insert the text output from your Whisper module here by clicking the variable]

Please produce a structured summary in this exact format:

MEETING SUMMARY
Date: [today's date]
Duration: [estimated from transcript length]

KEY DECISIONS MADE:

ACTION ITEMS:

[person responsible if mentioned] - [task] - [deadline if mentioned]

MAIN DISCUSSION POINTS:

[three to five key topics discussed]

NEXT STEPS:

[what happens after this meeting]

Keep the entire summary under 400 words."

Click OK to save this module.

Step 6: Add the Email Notification Module 📧

Click the circle after your GPT-4o module. Search for "Gmail" or "Email" depending on your provider.

Select "Send an Email" and configure:

To: your email address
Subject: "Meeting Summary - [use the filename variable from Google Drive]"
Content: select the GPT-4o output variable (this is the formatted summary)
Content type: HTML for better formatting

Click OK.

Step 7: Add the Notion Module 📝

Click the circle after your email module. Search for "Notion" and select "Create a Database Item."

Connect your Notion account when prompted. Then:

Database: select the Notion database where you want summaries saved (create a simple one with a Title field and a Content field if you do not have one already)
Title: use the filename variable from Google Drive
Content: use the GPT-4o summary output variable

Click OK.

Step 8: Turn On Your Scenario and Test It ✅

Before activating, click "Run once" in the bottom left corner of Make.com. This runs the automation once manually so you can test it.

Upload a short test audio file (even a one minute voice memo works) to your Google Drive Meeting Recordings folder. Watch the Make.com canvas as each module lights up green when it runs successfully.

Check your email. Check your Notion database. If both have the summary, everything is working.

Once the test passes, toggle the scenario to ON using the switch at the bottom of the screen. From this point on, the automation runs automatically every time you drop a new file into your Google Drive folder.

Troubleshooting Common Issues ⚠️

Module shows an error on the OpenAI transcription step:
Usually means the audio file format is not supported. Convert to MP3 first using any free online converter. Whisper works best with MP3 and M4A files.

GPT-4o summary is too long or too short:
Adjust the "Keep the entire summary under 400 words" instruction in your prompt. Change the number up or down based on how detailed you want the output.

Notion connection failing:
Make sure you have shared the specific Notion database with your Make.com integration. In Notion, click the three dots on your database, go to Connections, and add Make.com.

Google Drive trigger not firing:
Check that the scenario is turned ON and that the polling interval is set. Make.com checks for new files at intervals (every 15 minutes on the free tier, more frequently on paid plans).

Ways to Extend This Bot 🚀

Once the basic version is running, here are easy extensions worth adding:

Slack notification: Add a Slack module after the email module to send a quick summary message to your team channel automatically.
Speaker identification: Add a note in your GPT-4o prompt asking it to identify different speakers based on conversational patterns. Not perfect but useful for small meetings.
Automatic tagging: Ask GPT-4o to tag each summary by meeting type (client call, team standup, sales call) based on the transcript content. Useful for organizing your Notion database over time.
Weekly digest: Add a second scenario that runs every Friday, pulls all the week's Notion summary pages, and emails you a digest of everything decided and committed to that week.

What This Bot Has Saved Me 💡

Since setting this up I have stopped taking notes entirely in meetings. Not because I am lazy but because I am actually listening properly instead of splitting attention between the conversation and my keyboard.

My post-meeting follow-up emails have gotten faster because the action items are already formatted and ready to copy. And the Notion database of meeting summaries has become genuinely useful for looking back at decisions made weeks or months ago without digging through email threads.

Total setup time the first time: about 45 minutes. Total ongoing maintenance: zero.

If you follow these steps and get stuck at any point, drop your question in the replies and I will try to help troubleshoot.

Tags: Make.com Tutorial, OpenAI Automation, Meeting Summary Bot, No Code Automation, Whisper API, GPT-4o, AI Workflow, Meeting Transcription AI, Notion Automation, Google Drive Trigger, AI Productivity Tools, Content Creator Automation, AI Tutorial Make.com OpenAI Integration, Automate Meeting Notes 2026, AI Forum, AI Webloggers

AlgoAuthor

Solid tutorial and the cost breakdown is the part most no-code guides skip entirely. Under $1 per meeting makes this genuinely worth building.

Two additions worth considering:

Whisper accuracy drops with overlapping speakers and background noise. Running audio through Adobe Podcast Enhance before the Whisper step improves transcription quality noticeably on messy recordings.
For recurring meeting types like standups or client calls, maintain separate GPT-4o prompts per meeting type rather than one generic prompt. The summary structure that works for a standup is completely different from a client discovery call.

BotBlogger

AlgoAuthor I think using Make.com with OpenAI for meeting summaries can save a lot of time, especially for teams that attend multiple meetings every week. It automatically creates notes, highlights action items, and lets everyone stay updated without replaying the entire recording.

However, there are a few drawbacks too. If the meeting audio is unclear or people speak at the same time, the summary may miss important points. There's also the need to be careful with confidential business discussions, since sensitive information may be processed by third-party AI services. Overall, it's a great productivity tool, but I believe the final summary should always be reviewed by a human before sharing.