It’s probably hard to count how many times you’ve heard someone say that “generative AI has changed everything.” Browse the headlines on virtually any news organization’s website and you’re sure to come across something declaring that work as we know it has fundamentally changed.

I suppose in some ways, it has. But I think the real reason we see headlines like that is because it simply feels like it has. And there’s a world of difference between “feels like” and “is.”

If you subscribe to a frontier model these days—or, heck, even if you don’t and just use the free version—you now have infinite access to information, ideas, and answers. Any time the muse hits, you can prompt your way into insights that used to take hours or days, and it feels like a superpower, or a cheat code—that is, until you realize you’ve spent an hour talking about your problem instead of doing something to actually solve it. It feels productive because the conversation is smart and the ideas are good.

But when you close the tab to your browser and walk away from your computer, nothing has actually changed.

The Feeling of Progress

This is yet another maddening paradox of AI: it feels like we’re making progress when we’re not actually doing anything at all. Information is instant. Feedback is endless. And the result is a kind of motionless productivity—a loop where we collect more insights, more frameworks, more possibilities, and then mistake all of that for actual work, when the reality is we haven’t taken any action at all.

AI might be the world’s most powerful technology, but it’s only the world’s most powerful technology if you use it to actually do something. Your intent may be to solve a problem. But your intent only matters if it’s paired with action.

Redefining AI: Action + Intent

In the AI consulting world, lots of people have been redefining what the letters AI stand for. I’ve heard it stands for average intelligence, assume incorrect, or even amazing interns. So I figured I might as well jump in and come up with my own: AI = action + intent.

Action without intent is essentially chance. Intent without action is essentially thought. But when you put them both together, you start to realize the gains the headlines you’re reading are highlighting. AI stops being a novelty and instead becomes a driving force towards solving the problem you arrived to the tool with in the first place.

When you take action, something happens. It might be a good something, or it might be a bad something, but it’s something, and that something is information you desperately need, feedback you can use to recalibrate your approach.

AI can be a game changer, but the keyword here is “changer.” And if you’re just thinking, then you’re not changing. So in a effort to help you, here are three steps you can take to move from thought to action.

1. Engage AI as thought partner. Give an LLM all the context you have about a problem you’re facing using the CRIT approach to prompting. (Context, Role, Interview, Task)

2. Design the pilot. Work with the LLM to define a three-week pilot with an easy-to-understand success metric. (Time saved, money saved, quality increase, etc.)

3. Reflect and refine. Keep what works (or roll it into a larger-scale pilot), scrap what doesn’t, and document both.

After, you won’t just feel like you’ve done something, you will have actually done something. And that won’t just feel good. It will actually be good. Intent brings clarity. Action brings results. And every real breakthrough happens at that intersection.