Which AI Tool Should You Use? It Matters Less Than You Think

The most common question

It’s the most common question I get. “Should I use ChatGPT or Claude? What about Gemini? Is Copilot better for my business?”

I get why people ask it. There are dozens of AI tools out there, new ones every week, and everyone’s got an opinion. It feels like the most important decision you’ll make.

In practice, it’s one of the least important.

The tool is maybe 10% of the equation

Think about it like this. Give a professional chef a mid-range kitchen and they’ll make something incredible. Give anyone else the best kitchen in the world and the results will depend entirely on their skill, not the equipment.

The kitchen matters a bit. The skill matters enormously.

AI tools are the same. ChatGPT, Claude, Gemini, they’re all impressive. They all have strengths and weaknesses. But the difference between them is tiny compared to the difference between someone who knows how to direct AI well and someone who doesn’t.

I’ve seen people get better results from free tools than others get from expensive ones. The difference is always how they use it.

What actually matters

The skill that makes the real difference is knowing how to communicate with AI. Being specific about what you want. Giving it the right context. Leading the conversation instead of asking vague questions and hoping for the best.

This sounds simple, but it’s not something anyone does naturally. It’s a skill, and like any skill, it needs to be learned.

When someone asks me “which tool should I use?” I always ask the same thing back. “Can you describe, in detail, what you want it to do?” If the answer is vague, the tool choice is irrelevant. No tool is going to give you good output from poor input.

The subscription graveyard

Here’s what usually happens. Someone signs up for an AI tool, plays with it for an afternoon, gets mediocre results, and decides that tool isn’t good enough. So they try another one. Same thing happens. Then another one. Before long they’re paying for three or four subscriptions, none of which are delivering value.

The tool didn’t fail them. They never learned how to use it.

This isn’t a criticism. It’s genuinely what happens to most people, including engineers I’ve worked with. They assumed they could just type in a few words and get great results. When they didn’t, they blamed the model.

The fix is always the same. Learn the skill first, then pick the tool.

Pick one and get good at it

My honest advice? Pick whichever AI tool feels comfortable to you. ChatGPT is fine. Claude is fine. Gemini is fine. They all work well enough for the vast majority of business use cases.

Then spend your time learning how to actually use it. Learn to be specific. Learn to give context. Learn to break big tasks into small ones. Learn what AI is good at and what it isn’t. Understand that it’s trained on the internet and is just above average at most things, which means your guidance is what pushes the output from mediocre to genuinely useful.

Once you have that skill, switching between tools is trivial. You’ll know what to ask for, and any decent tool will deliver. Without that skill, the fanciest tool in the world won’t help you.

The returns aren’t even close

Every hour you spend reading reviews about which AI tool is 3% better than another is an hour you could have spent actually learning to use the one you’ve already got. The returns aren’t even comparable.

Get good at the fundamentals. The tool debate becomes irrelevant once you do.

If you’re not sure where to start with building that skill, that’s the foundation of everything I do.