A few weeks ago, Matt Shumer, CEO of OthersideAI, published an essay called "Something Big Is Happening." In it, he describes his experience watching AI go from useful tool to something else entirely—something that now does his job better than he does. He writes about telling an AI what to build, walking away for four hours, and coming back to find the finished product waiting for him. Not a draft; the whole finished thing.

Shumer's essay feels pretty urgent and might give you a small existential crisis. However, it's worth reading — not because I agree with every word (we differ on a few points), but because it shows you how the people building and using these tools every day are actually thinking about this moment. And it's a perspective that most business leaders simply haven't seen yet.

You Tried It. It Wasn't That Good. That Was a Lifetime Ago.

I keep having conversations with people who tried ChatGPT sometime in 2024, got a mediocre answer, and moved on. Naturally, they concluded that AI is “really just a better Google.” They might concede that it's a decent email writer or maybe even useful for brainstorming, but that’s about it. And, honestly, at the time, they weren't entirely wrong.

But that experience is now ancient history.

The models released in the last ninety days represent a fundamentally different class of technology. These aren't chatbots that answer questions. They reason through complex, multi-step problems. They write, test, and deploy code. They analyze data, synthesize strategy, and operate autonomously for hours. We've moved from a world where AI could help you draft a marketing campaign to one where it can build and run the campaign itself. AI agents—systems that don't just respond but take sustained, autonomous action—aren't a future concept. They're here. All you need to do is spend an afternoon with Claude Cowork and you’ll see what I mean.

If you're still thinking about AI as a glorified answer engine, you're evaluating the current landscape with an outdated map. And the distance between that map and reality is growing every week. Anyone who cites hallucinations, or who says they tried AI and it didn't work, is evaluating a technology that no longer exists. (We often say that the AI you use today is the worst AI you will ever use. AI is getting better by the day and the models we have access to are already one or two models behind what the AI labs are working on.)

The ceiling has moved so dramatically that the only honest way to understand it is to use it. And that's the core of the problem, because the leaders who most need to understand the shift are the ones who aren't engaging with it seriously.

AI Transformation Is Change Management

The most common misconception I encounter working with organizations is that developing an AI strategy is fundamentally a technology or IT initiative. It's understandable—after all, it involves technology, so it feels like it belongs in that box.

But AI transformation is actually an exercise in change management. And like all change management, it starts with people.

Every organization I work with—whether it's a large industrial company or a professional services firm—has the same pattern on its teams: a couple of people who are genuinely excited and who are already experimenting and trying to build with AI. Then a larger group in the middle who are curious but cautious, waiting for permission or direction. And finally, there’s always at least one person who simply will not engage at all, someone who has decided AI isn't for them and they’re “not going to train their replacements.”

That's fine and that's human. Not everyone is going to be excited about the technology, and not everyone needs to be an early adopter. But what can't happen is for the organization to stall because of it.

What makes this particularly hard for leaders is that you're trying to do two things at once. You're trying to operationalize AI across an organization—driving gains in efficiency, decision-making, and execution at scale. But to do that, you first have to empower individual learning. People have to understand how to use the technology themselves before you can unlock any of the organizational benefits.

This is where everyone gets stuck. They can see the gap between "I'm using ChatGPT to rewrite an email" and "AI is helping us run our business better." But they can't see the bridge. And so they stall. Which is something you simply can’t do anymore.

The Bridge Is Simpler Than You Think

Here's what I want you to hear: the bridge between individual use and organizational impact is not some massive IT deployment. It's governance, direction, and permission.

You don't need to spend months developing a comprehensive AI policy. You need the basics: clear guidelines on where AI can and can't be used, what data is off limits, and what "responsible use" looks like in your context. It could be more than that of course, and it probably should be if you're a larger organization, but it could also be a one-page document. You really just need to start somewhere.

Then you need to figure out where your people are and meet them there. Give them two things: permission to experiment and a little bit of direction on where to start. Every role in your organization—from admin assistants to salespeople negotiating deal points to legal teams reviewing contracts—can be enhanced by AI. But people need to discover that for themselves, in their own work, with their own problems.

When individuals start finding use cases that work—in their specific roles, with their specific challenges—you also need to operationalize that knowledge. You share what's working. You let one person's breakthrough become another person's starting point. That's how individual learning becomes organizational capability. That's the strategy.

Start Today

The organizations that figure this out first are going to pull away. Not because they have better technology, but because they have better organizational capability. The companies that don't will find themselves competing against organizations that are simply faster, sharper, and more efficient at nearly everything they do.

These aren't companies that need to become tech companies. They're companies that need to become companies that use AI well. There's an enormous difference. And the barrier to entry is far lower than most leaders assume.

You don't need a massive budget. You don't need a team of data scientists. You don't need to understand how the models work under the hood. You need to start using them seriously, put basic guardrails in place, and give your people the permission and direction to do the same.

The ground has shifted. It's more accessible than you think. And the time to start is today.