You're not behind on AI. You're leaving money on the table.

There is a particular anxiety that shows up in conversations with business owners right now. It sounds like everyone else has figured this out and I haven't started. It is almost always wrong, and it is expensive, because it sends people looking for the wrong thing.

You are probably not behind. What you probably are is using AI without it being worth anything — and there is a large, boring, entirely fixable difference between those two situations.

Fifty-eight percent. If more than half of small businesses are using it, being a non-user would indeed put you behind. But hold that number next to a different one, measured a completely different way.

These two figures come from different studies measuring different populations, so treat the arithmetic loosely — but the direction is unmistakable. A lot of businesses say they use AI. Far fewer have put real money behind it. Which tells you what most "we use AI" actually means: somebody on the team opens a chatbot when they get stuck.

That is genuinely useful. It is also completely invisible in the business. It leaves no trace in your calendar, your close rate, or your payroll, because it never touched a process — it just made one person's afternoon slightly easier.

Using AI is a personal habit. Putting AI in a process is a business decision. Only one of them shows up in your numbers.

What "in the process" actually means

A process is something that happens whether or not anyone feels like it. The phone rings and someone answers. A quote goes out and someone follows up. An invoice arrives and someone types it in. These run every day, they have a cost, and — critically — they have a failure rate you are probably not measuring.

That failure rate is the money on the table. Not a hypothetical efficiency gain: demand you already paid to create, arriving at your business, and falling through a gap. Three of the most common:

The phone after five. Roughly 27% of inbound calls to home-services businesses go unanswered — that is one in four people who wanted to give you money, reaching nobody. Most of them do not leave a message. They call the next name on the list.

The quote nobody chased. A quote that gets a second touch closes at a meaningfully higher rate than one that gets none. Most small businesses have no systematic second touch, so the customers who were merely busy get quietly filed with the ones who said no.

The paperwork someone retypes. Invoices, intake forms, permits — arriving as PDFs and photos, keyed in by hand by someone whose time is worth considerably more than data entry.

None of those are AI problems. They are process problems that AI happens to be unusually good at closing, which is a different sentence and leads somewhere better.

The part almost everyone skips

Here is the most-quoted number in the industry, and it is worth understanding properly rather than being frightened by: about 95% of enterprise generative-AI pilots showed no measurable P&L impact. That is enterprise data — large organisations, formal pilots — and it does not describe a twelve-person contractor. But the mechanism behind it absolutely does.

Read the phrasing again: no measurable impact. A pilot that quietly saved four hours a week but was never instrumented counts as a failure in that number — not because it produced nothing, but because nobody could prove it did. We wrote about that at length in everyone is quoting the 95% stat wrong.

The lesson for a small business is not "AI does not work." It is that measurement is not the paperwork after the project — it is the first step of it. If you cannot say what a process costs you today, you will not be able to say whether anything you built helped, and you will end up in the same position as the enterprises: money spent, nothing provable, quiet embarrassment.

What it looks like when it works

Sixteen and a half hours a week, across the owner and staff, in the businesses where it lands. That is not a marginal gain — that is most of a full-time position, returned to a company that never had to hire one. And notice what it is not: it is not a smarter model or a bigger budget. It is the same technology everyone else already has, pointed at a process instead of a person.

The entry cost has collapsed too. Median first-month AI spend fell from about $50 for businesses that adopted in 2019 to about $20 for those adopting in 2024. Whatever is stopping this, it stopped being the price of the tools some time ago.

How to find yours

You do not need a strategy, a platform, or a transformation programme. You need one honest hour and a pen. Walk one customer all the way through your business — first contact, quote, job, invoice, the silence afterwards — and at every step ask three questions:

How many arrive here? How many make it through? What is each one worth?

Wherever the second number drops sharply and the third is large, you have found it. That is where AI should go first, and everything else can wait — including the parts that would be more fun to build.

The reason so few businesses do this is not laziness. It is that the person who knows the process best is the person running it, and they have no time to stand outside it and count. Which is most of what our work actually is: standing outside it and counting, then building the one thing the counting justified.

If you want that done on your business, the first call is free and the written report that comes out of it is yours to keep — see a full example of what that report looks like, or start one for your business.