Your first hour with AI, and not a dollar spent

The mistake almost everyone makes first

The first instinct is to go looking for the right tool. It is the wrong first move, and it is the reason so many small businesses have a subscription they stopped opening in week three.

Here is what happens. You sign up for something, you sit in front of a blank box, and you ask it a general question to see what it does. It gives you a competent, slightly bland answer. You think huh, neat, close the tab, and never build a habit. Nothing was wrong with the tool. The problem was that you gave it an audition instead of a job.

The fix is to decide what the job is before you open anything.

Pick the job first, and pick a boring one

You are looking for a task with four properties. Write down anything in your week that has all four:

• It repeats. At least a few times a week. Frequency matters far more than how long it takes once.
• It is mostly words. Writing, rewriting, summarising, formatting, pulling information out of a message and putting it somewhere else.
• You would recognise a bad result instantly. This is the important one, and we will come back to it.
• Nothing terrible happens if the first few attempts are wrong. Not payroll. Not anything that goes straight to a customer unread.

Good candidates in almost every business: turning your scribbled notes from a call into a tidy summary. Rewriting the same explanation for the fifth customer this month. Reading a long email thread and telling you what was actually agreed. Turning a list of jobs into a message you can send a client. Drafting the reply you have written two hundred times.

Bad candidates for hour one: anything with a number in it that has to be right, anything legal, anything that goes out under your name without you reading it.

What the hour actually looks like

Set a timer. Open the free tier of any of the mainstream assistants. At this stage they are close enough that the choice genuinely does not matter, and a free account is enough to learn everything in this guide.

Minutes 0–10: give it the context you would give a new hire

The single biggest difference between people who find this useless and people who find it indispensable is how much context they hand over. Most people type one line. Try this instead, in one message:

• Who you are and what the business does, in two sentences.
• Who the output is for.
• What you want, specifically.
• An example of a good one that you wrote yourself.

That last item does most of the work. Paste in something you wrote that you were happy with, say match this tone, and the difference is not subtle.

Minutes 10–30: do the real job, not a test

Take a real piece of work from this week, a genuine email or a genuine set of notes, and do it for real. Then compare what came back against what you would have written. Not is this impressive. The question is would I have sent this, after a small edit?

Minutes 30–45: push back once

Whatever came back, tell it what was wrong in plain language. Too formal. We do not say ‘reach out’. Cut it by half. Lead with the price. Then look at the second attempt. This round is the whole thing: a tool that cannot take a correction is not going to help you, and one that can just became a member of staff who never gets tired of being edited.

Minutes 45–60: write down what worked

When you finally get an output you would actually use, copy the instructions that produced it into a note on your phone or a document. That note is now an asset. Next week you paste it in again instead of starting over, and the week after you hand it to somebody else.

People who keep these notes get compounding value out of this. People who start from scratch every time quietly give up. That is nearly the entire difference between the two groups.

Three things that will waste your hour

Asking it for facts about the world. It will answer confidently and it will sometimes be wrong. For anything you would have looked up, whether a regulation, a specification, a date or a number, check it at the source. Treat everything it tells you about the outside world as a lead, not an answer.

Giving it a job you cannot check. Covered above, and worth repeating because it is the one people break.

Expecting it to know your business. It does not know your prices, your people, your suppliers or last Tuesday. It knows what you put in the message. Almost every disappointing result traces back to context that was in your head and never made it onto the screen.

What you will know at the end

You will have a clear answer to one question: is there a job in my week that this does well enough to be worth doing again next week? If yes, you have found your first real use and you should go do it twenty more times before you think about anything else. If no, you have lost an hour and saved yourself a subscription, which is also a good outcome.

Either way, you have not spent a penny, and you now have an opinion based on your own business rather than somebody else's demonstration.