New to this? Start here.
A four-part guide to getting trustworthy answers out of AI, no jargon, about 14 minutes end to end. By the last one you’ll have a simple check you can run on any answer that matters: a contract, a letter from a doctor, an email you’re about to send at eleven at night.
How to tell if an AI answer is true in 30 seconds
Two questions, and a prompt you can copy, that catch a confident, wrong AI answer before you act on it. No jargon, no AI knowledge needed.
Read part 1 →The whole method, in four questions
The four questions I run any AI answer through before I trust it, shown end to end on one everyday example, with a prompt you can copy. No jargon.
Read part 2 →Make it show its working
An AI hallucination is a model blending what it knows with what it invents, in one tone. One prompt sorts the two into separate lists, so you see the guesses.
Read part 3 →Where I actually use this
The two-question check works on any decision that matters. Here's where I push it hardest, and where I wrote down what AI got wrong as well as right.
Read part 4 →By the end you’ll run this on any answer before you trust it. Two questions, under a minute. Here it is on a phone contract you’ve pasted in for a plain summary.
- Name the exact thing. Which provider, which plan, the price printed at the top. If it says £29 when the page says £39, that’s someone else’s contract being described in a very reassuring voice.
- Hand me one fact I can check in under a minute. One date, one number, one name you can hold against the real thing yourself. A made-up figure gives itself away the second you look.
Fail either question and the whole answer goes. Not the wrong line, all of it. See the full check and the four questions that follow →
The whole method, in one place.
The four parts above are the short version. The Prompt Stack is the full method, with prompts you can copy. And /lessons is the running log of the answers that didn’t survive a check. The point of the site, really.
Read the full method →