How do I turn my call recordings into content automatically with AI?

Every call you record is raw content. Most people know they could paste a transcript into AI and get posts out of it. Almost nobody does, because it's a chore. The value is making it happen without you.

"You can do this manually right now, but nobody does it because it's a pain in the butt."

What does the automation do?

1. Record every call. Use an AI notetaker that auto-joins every meeting. If you have to remember to turn it on, you won't. You want every Zoom you're ever on recorded.
2. Send the transcript to an AI model. Claude is probably the best one for this.
3. Give it custom instructions. Who you are, your background, documents about your work, and the style you want.
4. Ask for specific outputs. My example: a blog post, four LinkedIn posts and three tweets from each call.
5. Deliver it and ping you. It lands in Notion or Slack with a note that says it's ready.

None of this is advanced AI. Putting it on autopilot is the whole point.

Why run it on every call, not just podcasts?

Podcasts are the obvious use. The part I find most interesting is every single call. What I want: every time I finish a Zoom with my notetaker attached, a Notion page appears with ten different types of content from the call, without me doing anything. Pull the transcript, run it through a custom GPT built for that output, send the result to a new Notion page.

Here's why that matters to a busy owner. My content time is 9:30. I open Notion and there's already a ton of good content sitting there, written the way I asked for. I don't have to do anything.

How do I start without building anything?

Do it by hand first. Keep all your transcripts somewhere easy to grab. One morning I pulled a podcast transcript, threw it in AI and asked for ten ideas from that talk I could write about. Then I started writing about the first interesting one. I'm not great at thinking up random content, so this matters a lot for me.

Video works the same way. I showed my assistant how to copy a screen recording's transcript into a custom GPT that returns titles, keywords, long-tail keywords, a description, full articles, ten quotes and timestamps. That packaging is exactly the kind of repetitive work you want off your plate so you can just record.

Strategy calls are a third use. Drop the transcript into Claude, tell it what you talked about and give it a concrete goal, like making at least $3,000 a month within four months. Then keep going back and forth. I do this all the time, and it works surprisingly well. For meetings in general, I just ask for an overview or a to-do list.

What does it cost?

The price of any one tool isn't crazy if you can do tons of stuff with it. The problem is stacking them. You can't pay $40 to $100 a month for multiples of these things, one per output. That's the case for one pipeline that handles every call instead of a separate app for each kind of content.

What comes after call transcripts?

Build logs. I set up an AI agent that reads my files about what I built the day before with Codex, picks out which things would be most interesting for content, and sends me a short daily newsletter about my own work. Mining call transcripts is something other people already talk about. Mining build logs for content ideas is the next step.

The other direction is getting interviewed. You don't create content from scratch: an AI agent interviews you, then creates content from your answers. I've thought that for a long time, and I'll admit I'm still not doing content the way I should.