Why AI Will Make Us Better People

Years ago I was an information architect on an ecommerce project. We spent weeks on designs, product pages, a complex configurator, category layouts, the whole experience. I remember hanging everything off to the dev team and feeling great about it. We asked if it all made sense every week, every time they told us things were going well. Eventually they demoed what they'd built, and it looked pretty good, except for one small detail:

There was no shopping cart. No way to check out. No way to actually buy anything.

We had designed a store with no cash register.

The dev team wasn't wrong. We never detailed a checkout flow, so they didn't build one. To every business and design person in the room it was so obvious it didn't need saying. It's an ecommerce site, of course there's a cart. "Obvious" is a dangerous word.

I spent the next thirty years learning that lesson over and over. Developers aren't being difficult when they build literally what you asked for. They're doing their job. The problem is almost always on the side of the person making the request. We assume too much. We skip context. We mistake the picture in our head for a shared understanding.

Here's why I'm telling you this: that exact skill, learning to close the gap between what you mean and what you say, is what separates people who get incredible results from AI and people who think it's useless.

Most people who dismiss AI aren't wrong that their results are bad. They're wrong about why. They go to ChatGPT or Claude and type "make me a marketing plan for my hot dog stand" and get back something generic and lifeless. Then they say AI is overhyped. Here's the thing though, go to any marketing expert with that little context and you'll get the same bland output. The tool isn't the problem. The ask is the problem.

Getting a lot out of AI means doing the work that most of us skip in every conversation we have. It means thinking about who you're actually talking to. Not just "an AI" but what kind of expertise you want it to bring. Tell it you want a marketing strategist with food cart experience and you'll get a different answer than if you let it guess. That tiny bit of effort, thinking about your audience before you open your mouth, changes everything. It changes AI results. It also changes every meeting, every email, every conversation with your kids.

It means providing context. Did you already start this hot dog cart or is this a dream? Why hot dogs? What do you already know about marketing? I don't need your life story, but if you're asking me to help you, giving me some background so I'm not shooting in the dark makes the conversation dramatically better. AI works the same way. So does your coworker, your spouse, and your friend who you just called asking for a favor.

It means being specific about what you're actually trying to do. Are you about to spend a hundred bucks on Facebook ads tonight, or are you planning a whole brand launch? These are wildly different things. Having clarity on what success looks like before you start is the kind of thing that sounds obvious but almost nobody does. Not with AI, not with their teams, not with anyone.

It means telling people what you actually want back from them. Are we having a conversation, or do you need a document you can hand to someone? A brainstorm partner and a deliverable are different things, and if you don't say which one you want, you'll get whichever one the other person guesses. AI will do the same, overthink or underthink your needs depending on what it assumes you want.

There's a pattern here, and it's not really about AI.

Every skill that makes you better at prompting a large language model is a skill that makes you better at talking to people. Thinking about who you're talking to. Giving them context. Being clear about what you want. Saying what success looks like. These aren't prompting tips. They're just communication, and most of us are worse at it than we think.

I spent thirty years learning this the hard way, one checkout-less shopping cart at a time. Now millions of people are sitting across from a system that will do exactly what they ask instantly, and for the first time they're discovering what developers have been too polite to say for decades: the problem was never on their side of the conversation.

As we spend more time learning to get the most out of our LLM friends, I'm convinced many of us will become better communicators with the people around us too.