The best first question to ask an AI — and why most people get it wrong
When people first sit down with an AI tool, the most common thing they type is something like: "What can you do?" or "Hello." And while there's nothing wrong with that, it tends to produce a rather generic reply — a long list of capabilities that doesn't really tell you much.
The better approach: start with something real
The single most effective way to get value from AI straight away is to bring it a real problem — something you're actually trying to do. Not a test. Not a vague question. A genuine task.
For example, instead of asking "What can you do?", try something like: "I need to write a polite letter to my landlord asking them to fix the boiler. Can you help me write it?"
Immediately, you'll see something useful happen. AI will ask you a question or two, produce a draft, and show you — in a concrete and personal way — what it can actually do.
Why this works better
AI is at its best when it has something specific to work with. The more context you give it — the situation, the tone you want, who it's for — the better the result will be. A real task gives it all of that naturally, without you having to think too hard about it.
A simple rule to remember
Think of AI like a very capable assistant who is brand new to the job. They're eager to help and can do almost anything — but they need you to tell them what's needed. The more clearly you describe the task, the better the result.
Want to try this for yourself?
Getting the most from AI is something Kevin can walk you through properly at one of his workshops or in a follow-up session after a free library talk. If you'd like to give it a go, just ask Kevin to show you.
See upcoming events