How to Use ChatGPT to Create Better Talking Head Videos

Over the last few months, you’ve probably noticed a shift on social media. More creators are putting the camera on themselves and simply talking. 

No cinematic B-roll. No trending audio. Just one person sharing an idea, a story, or a lesson they’ve learned. People are craving content that feels more human and less produced. 

Ironically, though, I also think talking head videos are some of the hardest pieces of content to make. When there’s nothing else on the screen, the only thing carrying the video is your ability to communicate an idea clearly. 

Here’s how I use AI to make that a lot easier.  

All My Talking Head Videos Start Here

Most people assume the hard part is getting comfortable on camera. I don’t think that’s true. Don’t get me wrong, getting comfortable on camera takes time, but I think the hardest part is figuring out what you actually want to say.

The camera is a terrible place to organize your thoughts. By the time you press record, you should already know exactly what point you’re trying to make. 

That’s why almost every talking head video I record starts in ChatGPT long before I ever open Instagram. Not because I want AI to write the script. Because I want somewhere to think. 

Most of the videos that have resonated with my audience lately haven’t even been entirely about real estate. 

They’ve been about:

• becoming a better father
• building a business
• trying to stay healthy while raising a family
• lessons I’ve learned from sales

Or something I noticed during my day that I thought other people might relate to.

Those experiences become the raw material.

I’ll open ChatGPT and dump everything I’m thinking into one conversation. Every opinion, every story, every observation, every half-finished thought. It’s usually a complete mess, and that’s exactly how I want it. At this stage, I’m not trying to sound polished. I’m just trying to get the idea out of my head and onto the page.

How AI Helps Me Turn Recordings into Content Ideas

Once everything is there, the real work begins. I don’t ask ChatGPT to write the video. I start asking it questions.

• What’s the actual lesson here?
• Am I trying to make three different points instead of one?
• Is there a stronger angle? 
• Am I making this more complicated than it needs to be?
• If someone who doesn’t know me watched this, would they still find it interesting?

That back-and-forth is where most of the value comes from. 

A lot of people use AI like a ghostwriter. I use it more like a sparring partner. I want it to challenge my thinking, point out weak arguments, and help me find the clearest version of what I’m already trying to say.

Verbal & Text Hooks

Once I’ve figured out the lesson, I spend more time on the first sentence than almost anything else.

The verbal hook is the promise you’re making to the viewer. If the first sentence doesn’t make someone curious enough to keep listening, it doesn’t really matter how good the rest of the script is.

After the script is finished, I switch gears and start thinking about the text hook.

I think this is where a lot of creators get confused because these are two completely different hooks serving two completely different purposes.

The text hook gets someone to stop scrolling. The verbal hook gets them to keep watching.

One of the biggest mistakes I see is treating the text on screen like a transcript of the first sentence you’re about to say. Or worse, an ambiguous or clever text that doesn’t immediately make sense. 

Instead, I ask myself a different question: 

If someone knew absolutely nothing about this video, what’s the clearest sentence I could put on the screen that would make them stop?

Over the last year, I’ve become convinced that clarity beats clever almost every time. If your hook is vague, overly creative, or trying too hard to sound profound, people don’t stick around long enough to hear the point you’re trying to make. 

I’ve found that the more obvious the value is, the better the content tends to perform.

My Recording Workflow

As for actually recording the video, my workflow is pretty simple. Personally, I use a front-facing teleprompter in the Edits App because it allows me to focus on delivering the idea instead of trying to remember every sentence. 

I know some people think that’s cheating. I don’t really care. It works for me. That said, you absolutely don’t need one.

Some of the best creators I know hit record, talk through the idea, restart the sentence whenever they mess up, and trim those mistakes out in editing. That’s a perfectly legitimate way to do it, and if anything, it’s probably the fastest way to become comfortable talking to a camera.

The truth is, your first talking head videos probably aren’t going to be very good. (Mine weren’t either.) You get better by recording another one, then another one after that.

The best talking head videos don’t come from people who have the perfect script. They come from people who have something genuine to say and have put in enough repetitions that it sounds like they’re having a conversation instead of giving a presentation.

For me, that’s the real value of AI. It doesn’t replace my voice. 

It gives me a place to find it before I ever press record.