I stopped asking AI for answers and started asking for frameworks — and suddenly it all clicked

It doesn't matter if you're using ChatGPT, Gemini, DeepSeek or Claude; if there’s one thing AI is more than happy to give you, it's advice. Whatever your problem is — a problem with your boss at work, how to deal with your eccentric uncle at Thanksgiving, or which fridge you should choose for your kitchen — AI loves to tell you what it thinks you should do.

It will even give you medical, legal and financial advice if you ask for it, which obviously you should take with a pinch of salt, because those decisions clearly need to be made by a professional.

The only problem is, a lot of the time, the advice doesn’t work. That’s not to say AI’s advice is bad… it's quite often very good. But one thing you’ll notice is that the process never ends. One problem soon gets replaced by another, and you’re back to talking to a chatbot at 2.00am, ruminating over your life choices when really you should be asleep.

AI always wants to please

There’s also the issue that, like an eager puppy, AI wants to please you, its owner, and when you ask it for advice, it tells you what it thinks you need to hear, not what’s good for you. It doesn’t always see the big picture.

If what you need is help with something simple and functional, like the correct wording to write an email to contest a parking fine (I’ve used it for this exact purpose this week, and it did a great job), that’s one thing, but if you’re asking ChatGPT whether you should break up with your partner, or quit your job, that’s something else entirely.

We also don’t want to become dependent on AI to solve all our problems, because research has shown that the more we do, the more we start to risk experiencing a particular type of burnout known as smoothout. What we need is AI to support us in making our own decisions, rather than make them for us, and that’s where the “framework” prompt can come in handy.

The Framework Prompt

Instead of describing a problem to your chatbot and asking “What should I do?”, try adding “Don’t give me an answer, give me the framework I need to make this decision myself” to the end of your prompts. I’ve done this with numerous problems now, and it forces you to actually work through what you think and feel, and it supports you in coming to your own decision.

Quite often, the AI will give you a 10-step framework to run on your problem. It starts by asking you to define the real problem, and it goes on from there with steps specific to what you’ve asked. It can be an illuminating process.

Don’t expect the framework prompt to work for simple questions, especially ones where there’s one answer, like “Who is the President of the United States?”, but for hard questions, it’s a great option.

Try it, and I promise you that if you work through all the steps, you’ll get a much clearer picture of what you need to do, and you can reach your own conclusions without the existential guilt of having handed an important decision over to a non-human AI.

You might not get an instant answer, but you’ll end up with something more useful — a decision you actually believe in.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.

The best computers for all budgets

➡️ Read our full guide to the best computers
1. Best Windows:
Dell Tower Plus
2. Best Mac:
Apple Mac mini M4
3. Best Mac AIO:
Apple iMac 24-inch (M4)