How much does AI actually cost for a small business?

This is the question I get within about four minutes of every first call, and most websites dodge it. So let me not.

The honest shape of it

The tools underneath are cheap. That’s the part that surprises people. The AI itself — the actual thinking bit — usually costs a few dollars a month for a small business’s volume. Sometimes less. The software that connects things together is often free at the size you’re working at.

So if the ingredients are cheap, what are you paying for?

You’re paying for someone to work out what to build, build it properly, and make sure it doesn’t quietly break in six weeks. That’s it. That’s the whole cost.

Why nobody gives you a price list

Because “AI automation” isn’t one thing, and anyone who quotes you before understanding your week is guessing.

Automating one monthly report is a small job. Automating your entire enquiry-to-invoice process is not. Same words, very different amounts of work. A price list would either overcharge the simple job or badly underestimate the complicated one.

What I can tell you is the shape: a single well-defined system — one report, one inbox, one posting schedule — is a low four-figure build for most small businesses, and then it costs very little to keep running. Bigger, multi-step systems go up from there.

The number that actually matters

Here’s the question worth asking instead of “what does it cost”: how many hours a month does this job eat, and what’s an hour of your time worth?

If you spend six hours a month on a report, that’s seventy-odd hours a year. Put your own number on those hours. Now compare it to a one-off build cost. The maths usually answers itself, and when it doesn’t, that’s genuinely useful information — it means don’t automate that one yet.

I’ve told people not to hire me on this basis. It costs me a job and saves them money, and I’d rather be the person who does that.

The cost people forget

Not money — attention.

A system that needs babysitting isn’t saving you anything. If you have to check it every day, log into something, remember a step, then you’ve swapped one job for a slightly smaller job. That’s a bad trade and it’s very common.

Good automation is boring. It runs, it tells you it ran, and you forget it exists. When you’re comparing quotes, ask what happens when it fails. If the answer is vague, the real cost is going to be higher than the number on the page.

So what should you do first

Pick the most annoying repeating job in your week. Not the biggest — the most annoying, because that’s the one you’ll actually notice disappearing.

Work out roughly what it costs you in time over a year. Then have a conversation with someone about what building it would take. If the numbers don’t work, don’t build it. If they do, you’ve got your starting point.

That conversation is free with me, and I’m fine telling you the numbers don’t work.