Stop Letting AI Tell You How to Work

Everywhere I turn, someone has advice on the “right” way to use the latest AI tool. Download this prompt library. Subscribe to that hack list. Install this plugin. Follow that workflow. It’s as if there’s a single approved route to making AI work, and if you’re not on it, you’re somehow doing it wrong.

But surely there is no right way?!

A calculator doesn’t decide which sums I should do; a keyboard doesn’t dictate what I should write; and AI is no different. Its real value is in how it fits into your work, your needs, your priorities.

The danger of all the “how-to” noise is that it subtly shifts the power dynamic. Instead of being a tool that serves us, AI starts to feel like a system we have to keep up with and we end up chasing features, integrations, and workflows that may have nothing to do with the problems we’re actually trying to solve.

Tools Should Adapt to Us, Not the Other Way Round

The most effective tools in history have been the ones that disappear into the background, helping us do what we already wanted to do, faster, better, or more easily:

• Spreadsheets didn’t tell finance teams what questions to ask. They just gave them a faster way to answer.

• Email didn’t tell us what to communicate. It just sped up the process of doing so.

• Word processors didn’t invent new genres of writing. They simply made editing easier.

AI is no different. It doesn’t need to prescribe new behaviours to have value. It needs to make the things we already want to do easier, clearer, faster, or more imaginative.

Be Selfish With AI

So what’s the best way to use an AI tool? Be selfish.

• Use it when it makes life easier. If it helps you draft, refine, or test ideas, let it.

• Ignore it when it doesn’t. If it complicates the task, walk away.

• Shape it to your needs. Don’t twist your work to showcase what the tool can do. Twist the tool to showcase what you can do.

For me, this means using AI as a strategic partner. Sometimes it’s a sounding board to test an argument. Sometimes it’s an editor that sharpens a draft. Sometimes it’s a researcher that gives me a fast overview of a new topic. And sometimes I don’t use it at all, because the job needs human presence, instinct, or empathy. That’s not a failure. That’s just good judgment.

Ignore the “Prompt Olympics”

One of the stranger trends around AI is the “prompt Olympics”, i.e., the constant competition to produce the cleverest, most elaborate prompts that wring maximum performance out of the tool. There’s a certain geeky fascination in this, but it misses the point. If I spend more time writing a prompt than doing the task, I’m not saving time, I’m losing it. Prompts should be practical. They don’t need to be perfect. And they don’t need to be the same as anyone else’s. The best prompt is the one that gets your job done.

AI as a Mirror, Not a Master

One of AI’s most useful qualities is that it reflects back what you put in. If you ask vague questions, you get vague answers. If you ask sharp, specific ones, you get sharper responses. In that sense, AI is more of a mirror than a master. It amplifies the clarity, structure, and intention you already bring; but that only works if you approach it with your own goals in mind. If you start with “what can the tool do?” instead of “what do I need?” then you risk letting the technology set the agenda.

Unlocking What You Can Do

In the end, AI isn’t about proving what the tool can do. It’s about unlocking what you can do. That might mean:

• Writing faster, so you have more time to think.

• Stress-testing an argument, so you walk into a meeting sharper.

• Exploring fresh angles, so your creativity doesn’t run dry.

• Delegating the drudge work, so you focus on the value-adding parts.

If it helps you do those things, it’s doing its job. If it doesn’t, then close the tab and move on.

The most powerful way to use AI is the simplest: on your own terms. Not to prove your mastery of prompts. Not to keep up with the latest hacks. Not to impress anyone else, but to serve your work, your voice, your purpose.

So the next time someone tells you you’re “not using ChatGPT properly,” remember: there’s no such thing. The only wrong way to use AI is to let it dictate what matters to you.