A simple AI system that turns call transcripts into website copy and emails

Many of us do countless hours of customer calls. But we never repurpose those calls.

Inside those calls are the things that actually matter: what customers struggle with, what they want, and how they describe it.

AI can now convert all of that into website copy, FAQs, email sequences, and sales insights — automatically.

Here’s how to build this system.

Here’s what to build

A system that:

1. Auto-collects Zoom/Meet transcripts into Google Drive
2. Cleans each transcript automatically
3. Summarizes it into structured insights
4. Extracts patterns across all transcripts (top pains, objections, keywords, etc.)
5. Automatically generates:
   • Website headlines
   • FAQ pages
   • Outbound emails
   • Objection-handling scripts
   • ICP profiles
6. And updates them every time a new call happens.

All you need is:

• Google Drive (free)
• Zapier or Make (free tier)
• ChatGPT Projects (works with free models)

Let’s build it.

Step 1 — Create your Google Drive workspace

In Google Drive, create a folder called Sales Engine.

Inside it create these 6 subfolders:

Sales Engine/  
raw/ 
cleaned/  
summaries/  
patterns/  
assets/  
database/

Open the database folder inside “Sales Engine.”
Create a Google Sheet in that folder named:

customer\_insights

We will use this later to track customer pains and patterns.

Step 2 — Send Zoom/Meet transcripts into “raw/” automatically

If you use Zoom

1. In Zoom settings → turn on Audio Transcript.
2. In Zapier:
   • Trigger: Zoom → New Recording
   • Action: Google Drive → Upload File → save to Sales Engine/raw.

If you use Google Meet

Meet already stores transcripts in Drive.

In Google Drive:

• Right-click the transcript → File automation → “Move file” → send to Sales Engine/raw.

Now your transcripts go to the same folder, every time.

Step 3 — Create a ChatGPT project

• Go to ChatGPT → Projects → New Project.
• Name it:Sales Transcript Engine
• Inside this project, you’ll create 3 automations (ChatGPT calls them “Tasks”).

This Project will run all the automations.

Step 4 — Connect the project to Google Drive

Inside your Project:

• Open Data Sources
• Add the Sales Engine folder
• Give the Project read and write access

Now ChatGPT can read files from the folder and save new ones.

Step 5 — Automation #1: Clean the transcript

Inside the Project → Click New Task

Task name: Clean Transcript

Trigger: When a new file appears in raw\_transcripts/

Task instructions:

Clean the transcript by
  - removing timestamps 
  - fixing obvious errors 
  - formatting speakers 
  - removing system notices
Save the cleaned version into cleaned/ 
Use the same filename.

What happens: ChatGPT continuously monitors the folder and automatically cleans each new Zoom or Meet transcript.

Step 6 — Automation #2: Summarize into sales insights

Create another Task.

Task name: Summarize Transcript

Trigger: When a file appears in cleaned/

Task instructions:

Create a summary with these sections: 
 - Top customer pains 
 - Objections 
 - Goals 
 - Exact quotes (copy the customer's words) 
 - Competitors mentioned 
Save the summary into summaries/ as a JSON file.

This all happens inside the ChatGPT Project.

What happens: Once the transcript is cleaned, this second automation produces a high-quality summary

Step 7 — Automation #3: Update your customer insight sheet

Create a new Task.

Task name: Update Insight Sheet

Trigger: When a file appears in summaries/

Task instructions:

Open the Google Sheet in /database/customer\_insights. 
Add each pain point, objection, and customer quote from the summary as a new row. 
If a pain point already exists, increase its "count" number. 
Save a file called patterns.json inside the patterns/ folder. 
This file should contain the top pains, objections, and common words customers use.

What happens: Your Sheet becomes a living, constantly updated picture that continually changes to show what people want.

Step 8 — Automation #4: Create sales assets automatically

Now that your Project has a patterns.json file, you can generate assets automatically.

Create three small Tasks in the same ChatGPT Project. Each Task triggers when patterns.json changes.

Website copy task

• Trigger: patterns/ file updated
• Action: Write headline, subheadline, 3 bullets
• Save to: assets/website\_copy.md

FAQ task

• Trigger: patterns/ file updated
• Action: Convert common objections into an FAQ
• Save to: assets/faq.md

Outbound email task

• Trigger: patterns/ file updated
• Action: Write 5-email outbound sequence
• Save to: assets/outbound\_sequence.md

All Tasks live and run inside your ChatGPT Project.

Step 9 — Turn on automation features in the project

Inside the ChatGPT Project:

• Folder watching: ON (so it notices new transcripts)
• Task chaining: ON (so tasks run in order: raw → cleaned → summary → patterns → assets)
• File output: ON (so the Project saves finished work back into Drive)

Now, the system runs by itself.