At one of my AI workshops last year, a personal chef showed up.
She told me upfront that she was probably the least technically-minded person in the room. No coding experience. Barely used anything beyond basic apps. Came because a friend dragged her along.
By the end of the afternoon she had built a custom GPT that automated her weekly menu-writing process. Something that used to take her four hours was down to thirty minutes.
She wasn’t an outlier. She was the rule.
The Problem With Explaining AI
Here’s the pattern I see constantly: someone reads articles about AI, watches YouTube videos, maybe even takes a course. They understand, in the abstract, that AI is powerful. They know it can help with things like writing, research and automation.
But when you ask them what they’d actually build for their work? Blank stare.
The knowledge is there. The application is missing.
This is the “you don’t know what you don’t know” problem. You can’t imagine use cases you’ve never seen. The mental models haven’t formed yet. All the explanation in the world doesn’t fill that gap — because the gap isn’t information, it’s experience.
I figured this out early in my workshop days. I started teaching AI workshops in Austin because my online courses weren’t creating the same lightbulb moments for local people. At the first few workshops, I showed up with slides. Explanations of what AI could do. Clear frameworks for thinking about it.
Good content. But not the right format.
The shift happened when I started designing workshops around live demos instead of explanations. Not “here’s what you could do with AI.” But “watch me do it right now, in real time, on a real problem.”
What Happens During a Live Demo
When I do a live screen-share demo with a client or in a workshop, I usually don’t get more than about five minutes before someone interrupts.
Not because anything went wrong. Because they just had an idea.
“Wait — could you do that for my client reports?”
“That same thing would work for my research workflow, right?”
“Is there a way to apply this to the emails I get from vendors?”
The ideas come fast because seeing a workflow execute on a real task makes the possibility space concrete. Before the demo, they knew AI could “help with work.” After the demo, they know it can run their specific meeting follow-up, their specific content pipeline, their specific data problem.
I had a coaching call with Ilias — a structural engineer and investor — where instead of answering his questions about AI, I just shared my screen and started running my agents. Within ten minutes he had generated more actionable ideas for his own workflow than in any conversation we’d had before.
He didn’t need to understand the mechanics. He needed to see the shape of it.
Why Most AI Teaching Fails
Most AI content — articles, courses, even workshops — leads with explanation. Here’s what a large language model is. Here’s how prompting works. Here’s a framework for thinking about automation.
All of that is useful eventually. But it’s almost never the right starting point.
People don’t get excited about how AI works. They get excited about what it makes possible for them. And they can’t know what that is until they’ve seen something close enough to their own reality to connect the dots.
The chef in my workshop didn’t care how GPT worked under the hood. She saw me generate a structured recipe format from a few bullet points, realized that was exactly what she spent hours doing every week, and the rest was just her building her own version.
That’s the moment. That’s what demonstration creates.
The Principle Applied
This applies beyond AI teaching. It’s a principle about how people learn anything that requires imagination.
If you’re trying to help someone see what’s possible with a new tool, system, or approach, explanation creates awareness. Demonstration creates belief. And you need belief before anyone does anything.
So if you’re teaching AI — to clients, to a team, in a workshop — here’s the practical shift:
- Instead of explaining what AI can do for them, show them what AI does for you.
Run your actual workflows. Share your real screen. Use your live environment.
The imperfection is part of it. When something takes an unexpected turn or you have to troubleshoot in real time, that’s not a flaw — that’s the proof that this is real and you know how to navigate it.
The polished demo reel doesn’t create the same response as watching someone actually use their tools.
- Structure demos around their context, not yours.
If you’re coaching a lawyer, find the part of your workflow that maps closest to contract review or client intake or research. If it’s a restaurant owner, find the part that looks like menu planning or scheduling or marketing.
The closer the demo is to their actual work, the faster the ideas flow.
- Ask: “What in your work looks like this?”
After a demo, that’s the only question you need. It invites them to do the work of connecting their problems to what they just watched. Most of the time they already have three answers before you finish the sentence.
The Bottom Line
You can explain AI for an hour and people will nod along. Show them a 45-second workflow and they’ll interrupt you with ideas.
The gap isn’t knowledge. It’s experience. And experience comes from seeing, not hearing.
If you want people to believe in what’s possible, stop explaining and start showing.
I run hands-on AI workshops in Austin and online where the entire structure is built around live demos. If you’re building your own AI training program or want to develop this skill, the Two Hour Workday workshop covers the same demo-first methodology.
