Last November, I was wrapping up a live masterclass on annual reviews and planning for 2026. One student shared something that stopped me mid-thought.
She said AI had basically reorganized her whole work life. She was socializing more, thinking more clearly, and letting automated systems handle the repetitive stuff. Different person at work, she said.
My response surprised even me: “Yeah, and most people still haven't started.”
She looked a little confused. We'd spent the last hour talking about AI like it was the obvious thing to do. And it is… for us. But step outside the circle of people who spend time building workflows and talking about agents, and the picture looks completely different.
The Adoption Gap Is Bigger Than You Think
Here's what I see every time I run an AI workshop in Austin:
I have somewhere between 3 and 6 professionals in every group who've never seriously used ChatGPT beyond a one-time curiosity search. Not people who tried it and quit. People who literally haven't started.
These are doctors, lawyers, real estate investors, and business owners. People with teams. People managing portfolios. Not behind by personality or intelligence — just not there yet.
And my workshops usually fill with the interested people. The ones who showed up.
Think about the ones who didn't.
The World Economic Forum estimated that AI adoption in daily professional work is still in the single digits as a percentage of the global workforce. We talk about AI like it's ambient now, like everyone has it. But most people are still watching from the sidelines, waiting to see if it sticks.
It has stuck. They just haven't caught up yet.
Three Levels, and Where Most People Actually Are
I teach a simple progression in my workshops. I call it the AI Fluency Levels:
Level 1 — AI Assisted: Using AI tools like ChatGPT or Claude to help with tasks. Writing, summarizing, researching. One prompt at a time, manually.
Level 2 — AI Workflows: Chaining tools together. Connecting your calendar to your notes, your email to a draft generator, your meeting transcripts to a weekly summary. Automated sequences that run without you clicking everything yourself.
Level 3 — Building Agents: Fully automated systems. AI agents that handle tasks end to end — scheduling, research, drafting, follow-up — while you review and approve.
Most of the world is at Level 0. Not Level 1. Zero. They've heard about AI. They've read headlines. But they haven't actually used it in a way that saved them time or changed their workflow.
If you're reliably at Level 1, you're in a small group. Level 2 puts you in a much smaller one. Level 3? You're genuinely rare.
What That Gap Actually Means for You
I want to be honest about something. The AI chatter online is loud. People posting about their agents, their setups, their workflows. It can feel like everyone is doing this already. Like you're just keeping up.
But that's the noise of the interested minority. The majority is still watching.
That creates what I'd call an arbitrage window. A period where the tools are available, reasonably priced, and powerful — but not yet widely used. The people who build fluency now will have a compounding advantage over the ones who wait.
I've seen this before with the internet itself, with smartphones, with social media. There's always a window before something becomes table stakes. Once it's table stakes, the edge is gone.
We're still inside that window with AI. But it's closing.
What I Actually Watch Change in People
The chef who came to my workshop in Austin last year is one of my favorite examples of this.
Zero technical background. She told me before we started that she was probably the least tech-savvy person in the room. She'd never built anything with software. She was sure she'd be lost.
By the end of the day, she was creating custom GPTs that automated parts of her work. Client onboarding emails, menu drafts, response templates. Built by her, for her business, in one day.
That's what happens when someone moves from Level 0 to Level 1 in a short, focused environment. It's not a years-long journey. It's a few hours of genuine engagement, and everything shifts.
She emailed me a week later and said she'd automated something else. And then something else after that. That's the compounding that happens once someone crosses the line from watching to doing.
The Part That Still Gets Me
I was talking with Michael Housman a few months back — he speaks to enterprise audiences 50 to 60 times a year on AI topics. Smart guy, genuinely plugged in to where adoption is and isn't happening.
We both agreed on the same thing: even the most digitally sophisticated organizations are just now figuring out how to use AI for basic productivity gains. Not agents. Not workflows. Just… prompting well.
And outside those organizations? The adoption rate is even lower.
I have clients right now who are running full automation stacks — email triage, meeting prep, research synthesis, CRM updates — and saving 20 to 40 hours a week. Their competitors aren't doing any of that. Same industry, same market, massively different output.
That gap gets bigger every month the other side waits.
So What Do You Do With This?
If you're already past Level 1, keep building. Don't wait for the tools to be perfect. They're good enough now.
If you're at Level 0 or 1, pick one workflow to improve this week. Not a big overhaul — just one thing. Automate one email response. Use AI to prep for one meeting. Get your first reliable win and build from there.
The one principle I keep coming back to is start small, iterate. Don't build the whole system at once. Prove a small win. Then expand. Every agent you add compounds the previous ones.
The forefront doesn't stay the forefront forever. But we're still at it. Might as well act like it.
Thanh Pham runs AI productivity workshops in Austin, TX. If you want to see what AI can actually do for your team, check out the Productivity Academy.
