How to Start Using AI at Work if You've Never Used It

A lot of people have already opened an AI tool, typed something, read the answer, thought "that's interesting," and never went back. It's not a lack of interest or ability. It's that nobody told them what to actually use it for in their own work. You don't learn AI by reading about AI — you learn it by using it on a task you already do and are already tired of.

The Mistake of Starting With the Hard Stuff

Most people start by asking for something big: a strategy, a business plan, a full analysis. The answer sounds good but is generic, and the conclusion is that "it doesn't work for my case." The problem is the starting point. Big tasks need context the tool doesn't have. Small, repetitive tasks work on the first try. And small, repetitive tasks are exactly where most of your wasted time lives.

Five Tasks to Try This Week

• Draft an uncomfortable email: a payment reminder, an apology, a rejection.
• Summarize a long document someone sent you that you haven't read.
• Turn your loose meeting notes into action items with owners assigned.
• Rewrite a text you already have to make it shorter and clearer.
• Prepare ten questions for an interview or a meeting with a supplier.

All five have something in common: you already know what a good result looks like. That's key at the start, because you can judge whether the answer is actually useful. When you ask for something you can't evaluate, you don't learn anything — you just accept whatever came out.

The Difference Between Asking Poorly and Asking Well

Asking poorly is "write me an email for a customer who hasn't paid." Asking well is "write an email for a customer who's fifteen days late, with whom we have a good relationship we want to keep; friendly but firm tone, five lines max, closing by proposing a specific payment date." The difference isn't technical, it's about clarity. You're giving it the same context you'd give someone new on your team. That's the whole trick.

If you can't explain the task to a new hire in three lines, you won't be able to explain it to AI either.

What to Realistically Expect in the First Two Weeks

The first week, it's going to feel like it takes the same amount of time. That's normal: you're learning to explain something you used to do automatically. The second week you start reusing what already worked, and that's when the time savings shows up. By the end of the first month, most people have three or four tasks they do in half the time. It's not a transformation, it's a small, sustained change. And that's exactly what you should be aiming for at the start.

From Personal Use to Team Use

The next leap happens when what worked for you gets shared. If you found a way to draft quotes in a third of the time, that shouldn't stay stuck in your head. The teams that move fast are the ones that save their working prompts in a shared place and improve them together. That's where AI training makes the difference between a few people experimenting on their own and a team that works differently.

You don't need to understand how a model works to benefit from it, just like you don't need to know about engines to drive. You need to practice on something you care about, correct it when the answer doesn't work, and save what does. Start tomorrow with the task you dread the most — that's the one best suited for it.