How we test
AI for the rest of us. Plain guides to getting everyday jobs done with an AI assistant, tested before they go up.
What a guide gives you
Each guide is one job: planning a week of lunches, reading a lease, writing a complaint, preparing for an appointment. It starts with a prompt you can copy, then the steps, what a good answer looks like, where AI tends to get this job wrong, and what not to paste in.
We run every prompt
Before a guide goes up we run its prompt in an AI assistant and check the answer. The guide says which assistant we used and when. We never claim to have tested a tool we didn't use. AI tools change often, so when a guide stops working the way it says, we update it.
Where AI goes wrong
AI assistants can sound confident and still be wrong: they make up facts, slip on maths, miss local rules and use outdated information. Every guide tells you what to check, and against what. For money, legal and health questions, our guides help you understand and prepare. They don't replace a professional, and they say who to ask.
Privacy first
Every guide says what to leave out. Never paste passwords, ID numbers, card or account numbers, or other people's private details into an AI assistant. Each assistant's maker explains what it keeps and how to turn history off; we link to their own pages.
What we don't write
No guides to cheating, deceiving or harming anyone: no fake reviews, no impersonation, no getting around rules that forbid AI, and no snooping on people.
When we get something wrong
If a guide is wrong or out of date, we fix it, and the date on the page shows when it was published.
