At a recent talk, I asked the audience how many had tried ChatGPT. Nearly every hand went up. Then I asked how many paid for a subscription. Two-thirds of the hands dropped. Finally, I asked how many had experimented with Claude or Gemini. Only a few remained.

It was one of the trends I’d seen repeatedly over the last two years—lots of people had experimented, but few had gone as far as committing to the twenty-dollar-a-month license.

That exchange said a lot. Curiosity is everywhere, but commitment is rare. And it reflects what I’ve been seeing for two years now: we’re standing in the middle of a once-in-a-generation disruption, yet most organizations are still at the starting line.

 An Imperative Shift

When it comes to AI, I sometimes wonder if I’ve gotten the whole thing wrong.

But it’s not because I’ve read something to make me think that. It’s simply because so many companies—even at this point, two years after the commercial launch of ChatGPT—are still hesitant to experiment, or dismissive of AI altogether.

While there are certainly an abundance of success stories to be found, the most impressive of which might be Moderna’s AI transformation—where AI was leveraged to accelerate mRNA research, streamline vaccine production, and support every area of the company—the reality I see is that many companies are still struggling to take the first step.

To some degree, I understand this. It’s been a long time since I’ve worked inside of a large corporation, but I still remember how hard it was to be a part of process-driven organizational changes. I’m empathetic to the challenge of trying to cognize, adopt, and implement a technology that inspires both excitement and existential dread.

But it’s still no excuse. Your organization needs to find a way to begin its AI transformation. And the sooner, the better.

Reasoning is Here

Part of why I say this is because at the frontier model level, we’ve already moved from System 1 thinking (pre-trained instinctual responses) to System 2 thinking (deliberate reasoning). In layman’s terms, this means that models can now pause, evaluate, and reason through decisions like a human would. To put it another way, they can take the time they need to "think."

Perhaps that doesn’t mean much to you right now, but the first time you experience a reasoning model stop and think for a few minutes before providing you with, say, a comprehensive analysis and strategic recommendations for your company’s planned expansion into the Asian market, you’ll start to view the future very, very differently.

To be clear, the idea that I’m pushing here—AI transformation—doesn't presuppose that AI takes over every department in your organization. Rather, it simply takes part—in small ways at first, and then in bigger and more substantial ways. And then, later still, in ways as vital and necessary as electricity and computers.

Even if the next version of ChatGPT never comes out, even if Gemini and Claude never make another model update, even if this is the most capable form of AI we ever get to (doubtful), everything is already different. For the first time in history, your entire team can do more with more.

With some training, a few guardrails, and executive support, your teams can make gains never before possible, in ways you won’t fully understand until you try.

Don’t wait.

Assemble your AI Council, create your AI Charter, and draft your AI Policy.

Then, begin your transformation.

Look at how companies like Moderna leveraged AI not just in research but across their entire operations to gain a significant competitive edge. Ask your peers what they're doing. Read a book or two. Find a place to get some training.

There's no point in waiting any longer. You’ve got nothing to lose and everything to gain.