You’re Probably Using AI Wrong—Here’s 3 Ways to Fix That

Right now, AI is dominating headlines, as if it’s the second coming of sliced bread. Everywhere you look, someone’s promising that artificial intelligence is going to revolutionize how we work, think, live, and maybe even fold laundry. But in reality? Most people are still using it like a fancy search engine. And that’s a mistake.

It’s not your fault. After more than 20 years of typing a few keywords into Google and expecting the right answer to pop out, treating AI the same way feels instinctive. But unlike your standard search bar, generative AI tools like ChatGPT are designed to function more like a well-informed colleague than a digital Rolodex.

That’s why Paul DeJarnatt the Vice President and Head of Digital at Novus, an independent media planning and buying agency, shared three small—but powerful—shifts that can help anyone get dramatically more value out of AI tools. These aren’t sci-fi breakthroughs. They’re simple changes that will make your day-to-day work smoother, faster, and more fun.

1. Treat it like a friend or neighbor who’s really good at one thing

Imagine your neighbor is a retired tax attorney. You wouldn’t knock on their door and yell, “Best retirement plan 2025!” and then walk away. You’d ask them a question, give some context, and follow up based on what they said. That’s exactly how you should interact with AI.

Instead of tossing in keywords, have a conversation. Tell it what you’re working on, what you already know, and what you need help with. Ask for suggestions. Question the responses. Provide feedback. The more context you give, the better the output will be.

And while AI is shockingly good at guessing what you meant, it’s not infallible. A little skepticism goes a long way—treat it like you would that very helpful, slightly overconfident friend. Use it to stretch your thinking, not to follow directions blindly.

2. Use it for data analysis and organization

Here’s where AI stops being theoretical and starts saving you real hours in the day. With the right prompts, AI can analyze and organize data faster than any intern—and in the long run, faster than any human, period.

 Let’s say you’re dealing with multiple data sets from different sources. Maybe you’ve got weekly ad performance data from one platform, daily numbers from another, and store-specific reports from a third. Sounds like a nightmare, right?

 Here’s a real-life example from the VP at Novus: They had exactly that situation. Three systems, three formats, none of them aligned. Instead of spending an afternoon wrangling spreadsheets, they told the AI exactly how they wanted the data organized—weekly rollups, consistent formatting, and labeled by store. It spits out a decent attempt, makes one mistake, and corrects itself on the second pass. Five minutes later, they had a reusable workflow that still saves time.

If you’re dealing with sensitive data, use a platform with appropriate privacy protections or a secure enterprise setup. But for general info, AI can be your super-organizer.

3. Ask it to self-diagnose

This one might blow your mind a little. Let’s say you’ve spent 10 frustrating minutes trying to get AI to do something right. You’ve rephrased your prompt four times and simplified your data, and you’re still not getting what you want. Now what?

Ask it what went wrong.

 Seriously—try something like, “Why do you think you’re not getting this right?” More often than not, it will tell you. Maybe it misunderstood the structure of your request. Maybe your data has inconsistencies. Maybe your instructions weren’t clear enough.

Even better? Sometimes it will propose solutions and fix itself in real time.

As the VP at Novus put it, “It’s like working with a smart, intuitive new employee. They may not get it right on the first pass, but they can look at their process, see where it went wrong, and course-correct in the moment.”

The bottom line: It’s about retraining yourself, not just the AI

These three tips might seem simple but powerful because they push us to think differently. In the early 2000s, the people who mastered Boolean search terms and clever keyword combos became internet wizards. Today, the new power users are the ones who learn how to interact with AI in a more human, nuanced way.

Yes, AI is a tool—but it’s one that gets better when you meet it halfway. Learn how to guide it, question it, and collaborate with it, and you’ll not only get better results today, but you’ll be positioning yourself for long-term growth in a world that increasingly values AI fluency.

In short, the future belongs to the people who know how to talk to robots like people.