There’s a version of the AI conversation that bores me stupid, and it goes like this: the robots are getting cleverer, so the clever move is to hand over more and more of your thinking to them.
I use AI every single day. It runs half my admin. And I’m telling you that’s backwards.
The businesses I see doing genuinely well with AI aren’t the ones with the best tools or a prompt library the size of a phone book. They’re the ones where a human stayed firmly in charge of the thinking and made the AI do the arguing, the checking and the heavy lifting underneath. Human first. Robot second.
Because AI agrees with you by default
Nobody puts this on the sales page, but out of the box, AI is a yes machine. You bring it an idea, it says lovely idea. You bring it a plan, it makes the plan prettier. You bring it a mistake, and it will cheerfully help you make that mistake at speed.
Think of it as the world’s smartest graduate. Brilliant, keen, and on day one it knows absolutely nothing about you, your customers or what you’re trying to do. You wouldn’t let a graduate that eager run your strategy unsupervised. Same rules apply here.
So the value isn’t in what it generates. It’s in what you make it question. Which is where my three prompts come in. I use these every single day and they’re the whole difference between a posh search engine and a sparring partner.
The three prompts
One: “Reframe how we’re seeing this problem.”
For when it’s agreeing you into a dead end. You know the feeling, you’ve been circling the same issue for twenty minutes and the answers keep getting more polished and less useful. What you need isn’t a better answer, it’s a different question. This prompt makes it stop decorating your frame and challenge it instead. Honestly, half the time the reframe IS the answer.
Two: “Ask me clarifying questions until you’re 95% sure you can do this without a mistake.”
This one changed how I work more than anything else I’ve learnt. Instead of guessing what you meant and producing something confidently wrong, it interviews you. And the sneaky bonus is that answering its questions forces you to think properly about what you’re actually asking for. Most of the quality jump comes from your own answers. Don’t tell it I said that.
Three: “What are the gaps, risks or things we haven’t considered?”
The question your brain skips when it’s tired. Which, if you’re running a business around a family, is most of the time, let’s be honest. Ask it at the end of any piece of work or any big decision. Costs you one sentence. It has saved me from shipping half-finished thinking more times than I’d like to admit.
Look at what none of those prompts do
None of them ask the AI to think for me. Every one of them makes it help me think better. Challenge my frame, question my brief, attack my plan. The judgement stays mine. I’m just borrowing a tireless brain to do the arguing my tired one can’t manage at 9pm.
That’s what human-led actually means. Not typing nicely. Not whatever “prompt engineering” is supposed to be this week. Staying the decision-maker.
The women I work with don’t need artificial intelligence, they’ve got plenty of the real kind. What they need is somewhere to put their thinking that pushes back, holds the detail, and doesn’t need them to arrive already sorted.
One honest caveat
There are questions AI can’t answer no matter how well you prompt it, because it only knows what you can see. And when you’re too close to your own business, which is all of us, you need a human standing far enough back to spot what you can’t.
That’s the seat I sit in. I Can See Clearly Now is two hours, me in your business, asking the reframing questions and the gaps-and-risks questions, with the added bonus of having actually lived it.