Over the last year, AI has become much less of a writing tool for me and much more of a thinking partner. It helps me organize ideas, analyze information, and eliminate repetitive tasks that used to slow me down.
More importantly, it has shortened the distance between having an idea and publishing it.
The biggest misconception I see is that people expect AI to do all the thinking for them. They’ll open ChatGPT, type a sentence or two, and hope it magically produces something great. In my experience, that’s backwards. The quality of the output has far less to do with the prompt than it does with the information you give it to work with.
These are the workflows that have had the biggest impact on my marketing, and every one of them is something you can start using today.
I Stopped Waiting for Transcripts
One of the simplest changes I’ve made has probably saved me more time than any prompt I’ve ever written.
Whenever I record a talking-head video, appear on a podcast, or film an interview, I don’t wait for someone to transcribe it anymore. Instead, I play the video from my phone into ChatGPT on my computer and let it generate a transcript almost instantly.
Within minutes, I have something to work with.
From there, I can ask AI to write a stronger hook, identify the most interesting takeaway, turn it into an email, create an Instagram caption, or even build the foundation for an article like this one.
The faster you can convert spoken ideas into text, the faster you can improve them.
I Show AI Before I Ask AI
This is probably the biggest mindset shift I’ve made over the last year. Most people tell AI what they want.
I show it.
When I was trying to improve our headline-style real estate news graphics, I didn’t ask ChatGPT to write a better headline. I uploaded screenshots from TMZ, Worldstar, and some of our own highest-performing posts. Then I started asking questions.
- Why do these headlines make people stop scrolling?
- What patterns do you notice?
- What makes one headline more compelling than another?
The same workflow applies to almost everything I create.
If I’m building a carousel, I’ll upload examples that performed well.
If I’m writing a listing description, I’ll upload photos, property details, and notes from my walkthrough.
If I’m trying to improve a graphic, I’ll show AI multiple versions instead of describing them.
One of the biggest lessons I’ve learned is that AI performs dramatically better when it has something to observe before it has something to create.
Stop describing what you want. Start showing it.
I Turn Articles Into Content People Actually Want to Share
One of my favorite content series is our headline-style real estate news posts.
Whenever I come across an interesting article, I read it first without AI. I want to understand the story before I ask anyone or anything else to summarize it.
Then I’ll paste the article into ChatGPT and start working through it together.
- What’s the most surprising statistic?
- What’s the detail people are most likely to tell a friend about?
- If this had to fit on one slide, what absolutely has to stay?
Sometimes the biggest number in the article isn’t the best headline. Sometimes a smaller statistic creates much more curiosity. Other times, the story isn’t the data at all. It’s what the data means for buyers, sellers, or agents.
AI helps me find those angles faster, but the judgment still comes from me. That’s an important distinction.
I Almost Never Type Anymore
I almost exclusively use voice mode on ChatGPT. I think better when I talk.
I’ll use it while driving between appointments, going on a light jog, or immediately after finishing a meeting while the ideas are still fresh. Instead of waiting until I’m back at my desk and trying to remember everything, I’ll spend five or ten minutes talking through what happened.
Those conversations become captions, articles, videos, listing strategies, email newsletters, and sometimes ideas I never would have found if I had waited until later.
The blank page has become much less intimidating because I’m no longer starting from one.
I Ask AI to Challenge Me
This might be the workflow that’s helped me improve the most. I rarely ask AI if something is good. Instead, I’ll ask questions like:
- What would make someone stop reading here?
- What’s weak about this hook?
- Which argument would you disagree with?
- Where am I being too vague?
- How could this be more specific?
Most people use AI to validate their ideas. I use it to pressure test mine.
Some of my best-performing content didn’t come from AI giving me the answer. It came from AI pointing out the weaknesses I couldn’t see because I was too close to the work.
AI Hasn’t Replaced My Thinking. It’s Made My Thinking More Valuable.
Whenever people ask whether AI is going to replace marketers, agents, or creators, I think they’re asking the wrong question.
The best marketing has never come from typing words into a computer. It comes from experience, judgment, curiosity, and understanding what people actually care about.
AI can’t replace those things. What it can do is remove the friction that keeps great ideas from becoming finished work.
Every workflow in this article starts with something real. A conversation. A video. An interview. An article. A listing. A collection of screenshots. A voice note from the car.
The more context I give AI, the more useful it becomes.
That’s probably the biggest lesson I’ve learned over the last year. I don’t spend my time chasing better prompts anymore. I spend my time giving AI better inputs.
In my experience, that’s where the real leverage begins.
