A Small Business's First Thirty Days With AI

Most advice about getting started with AI is written by people selling something. Here is a month that costs about twenty pounds and tells you honestly whether this is worth more of your time.

Week one: watch, do not buy

Do not install anything. Do not sign up for anything beyond one paid assistant subscription, which is roughly twenty pounds.

Instead, keep a note on your phone. Every time you do something that feels repetitive, slow or irritating, write one line. What it was, roughly how long, and how often it happens.

That is the whole task. Most people are shocked by the list after five days, because these jobs are invisible while you are in them.

Week two: try the boring things yourself

Take the three most common items from your list. Sit down with the assistant for an afternoon and try to do them faster.

Drafting a quote. Writing a follow up. Turning notes into something a customer can read. Summarising a long email chain into what actually needs deciding.

You will find some of it works immediately and some of it is more trouble than doing it yourself. Both are useful findings, and you have spent nothing but an afternoon.

Be honest here. The temptation is to conclude either that it does everything or that it is useless. Neither is true.

Week three: pick one job and time it properly

From what you have learned, pick the single job that is most repetitive, most frequent, and most annoying.

Then measure it properly for a week. How long does it take each time? How many times does it happen? Multiply out to a year and put a value on your hour.

That number is the whole business case. If the job costs you eight thousand a year and a build costs twelve hundred, it is obvious. If it costs four hundred a year, leave it alone.

Week four: decide, honestly

You now know three things you did not a month ago. What is actually eating your week. Which parts you can handle yourself. And what one specific job is worth in real money.

At this point one of three things is true.

You can do it yourself. Plenty of jobs are a subscription and a habit. If so, you have saved yourself several thousand pounds and you do not need anyone, including me.

It needs building, and it is worth it. You have a number, a job, and a case. Now you can talk to someone and know whether their answer makes sense.

Nothing is worth it yet. Also a real outcome. If your business is small enough or your admin light enough, doing it by hand is correct. Come back in a year.

What to avoid in the first month

Buying a platform. Anything with a per seat licence and a dashboard is priced for companies far larger than yours.

Automating something broken. If customers are not buying, doing your admin faster will not help. Fix the offer first.

Doing five things at once. The first job teaches you the most. Do it properly, live with it, then decide about the second.

Taking advice from anyone who has not asked about your business. Including anything you read online, this included.

What good looks like at day thirty

You are not looking for a transformed business. You are looking for one specific job that used to take an hour and now takes ten minutes, and a clear sense of what the next one would be.

That is a good month. Anyone promising more in thirty days is selling.

If you want to go deeper on choosing the first job, what to automate first covers it, and how long until it actually saves you time is the honest timeline.

If you get to day thirty and want a second opinion on what you found, get in touch. Or see what I build.