What 120 CEOs Asked After the AI Demo
In November 2025, I ran a live AI demo for about 120 CEOs at the YPO AI Advantage Summit in Los Angeles. The demo worked. The questions afterward weren't about AI -- they were about creativity, their children, whether they'd started too late. Here are the three that kept coming back, and the working answers I've built since.
The format was hands-on -- Gamma for presentations, ChatGPT for real-time generation, image tools for visual concepts, every executive building from their own prompts. Standing in that room -- a Hong Kong trainer with a CUHK history degree, no startup to sell, first international keynote at this scale -- the distance felt measurable. But something showed up during the build that I now mention in every corporate session: the executives who got the best output weren't the most technical. They were the most specific about what they wanted. That observation anchors the first question.
"If AI can do this in minutes, what's left for creativity to mean?"
More than before, not less. What moved to the machine in that room was production -- layout, first-draft copy, image rendering. What stayed on the human side was judgment: knowing what good looks like, knowing who you're making it for, knowing when the output is almost right and what's wrong with it. The person who typed "make me a presentation about sustainability" got something generic and blamed the tool. The person who described a specific audience, a specific problem, and a specific tone got something they could use. Creativity didn't shrink. The production layer got cheaper, which makes the judgment layer -- the part humans keep -- more valuable, not less.
"Should my son still bother learning design?"
One executive asked this about his teenage son -- whether the skills a father had spent twenty years building were about to be worth less than a monthly subscription. I've heard versions of this question from parents in Hong Kong school halls since.
The execution layer of design -- layout, rendering, mechanical production -- is compressing fast. But the person who can look at five AI-generated options and know which one works for this audience, at this moment, in this market -- that person is more needed now than before the tools existed. What changes is what "learning design" means. Less time on production mechanics, more on developing the judgment that no tool replicates. The son should learn design. He should also learn to describe what he wants precisely enough that a tool handles the mechanical parts. Both skills make each other more useful.
"Is it too late for someone my age?"
The most common version of AI anxiety I hear, from boardrooms to PTA sessions. The tools in mid-2026 are meaningfully easier than a year ago, and they'll be easier again in six months. Starting now with one real task you already do weekly -- handled partly by a tool -- beats having started early and stopped. The variable isn't age or technical background. It's whether the first use case is relevant to your actual work.
There was no follow-up engagement from the summit -- no batch two, no department rollout. A single event, produced well, then over. The engagements I value most are the repeat bookings, where a company calls back months later with a different department. That's adoption. But something from this particular session changed how I work: when an executive asks about their kid or their own relevance, the question deserves a direct answer, not a pivot back to the tool. That took me longer to learn than it should have.
