Booking an AI Keynote Speaker? Here's the Talk That Actually Works

I get the same email three or four times a month now.

It opens with kind things about the TEDx talk. It moves into the event, the audience, the company, the date. And then it ends with what they want me to talk about, which is almost never what I do.

Most of the time the ask is some version of: come tell our people how to use AI. Show them the prompts. Make them excited. Get them off the fence.

I don't do that talk. I want to explain why, because I think the gap between what companies are asking for and what they actually need is the thing that's costing them right now.

Here is the talk that's in demand. Forty-five minutes. Lots of demos. Lots of "here's a prompt you can steal." The audience leaves saying that was great. They go back to their desks. They open ChatGPT. They paste the prompt. It works once. The next week they're back to where they were, which is, if I'm being honest, where most of corporate America is with AI right now: stuck in the experimentation layer, not the integration layer, not the strategy layer, not the layer that actually changes the business.

That talk feels good. It does not move anything.

The talk I give is different. It opens with an emergency room. It moves through a piece of software my team built called Lighthouse that watches what AI says about brands. It lands on a framework I call calibrated trust, which is three questions you ask before you act on anything an AI tells you. By the end the audience is quieter than when it started. Not because it was a downer. Because it gave them something to do that doesn't fit on a sticky note.

The people who book the first talk want their team to feel ready. The people who book the second talk want their team to be ready. Those are different things. I do the second one.

A few specific kinds of rooms where this lands:

Marketing leadership offsites. The CMO knows the team is using AI. The team thinks the CMO doesn't get it. Neither side knows what their actual exposure is. The talk surfaces it.

Executive briefings at companies that have already done the "AI 101" lunch-and-learn six months ago and want to know what comes after. (What comes after is harder. That's what I'm there to make the case for.)

Industry conferences where the lineup is already heavy on technologists, and the organizer wants someone who can speak to what AI is doing to the customer, the audience, the audience's parents, the audience's kids. The cultural side of the technical question.

University and student-leadership events. I do a version of the talk for people who are going to inherit this whole mess from us. It's the version I'm most careful with.

If you're trying to figure out whether your event is the right fit, here's the test I use. If your audience leaves the talk with a tool they didn't have before — a question, a framework, a way of looking at something they were already doing — it was the right talk for the right room.

If they leave with a prompt, I wasn't the right speaker. There are people who do that very well. I'm not one of them, and I'd rather tell you that now than show up and disappoint your CEO.

To book the talk, the contact form is the fastest path. Tell me the audience, the size, the date, and the version of the question you're hoping the talk answers. I read all of these. I write back the same day when I can.

The TEDx talk is online. If you want to know what the long-form version sounds like before you book, watch that first. It's the truest preview I can offer.

‍