The question, asked honestly

You have an idea. Maybe two. And somewhere between the spreadsheet and the license application, the question walks in and sits down: will AI eat this business before I get it off the ground?

You already know the two standard answers, because they are the only two the internet sells. The first is panic. Every business is about to be disrupted, act now, buy my course on surviving the apocalypse. The second is denial. AI can't do what a real tradesman does, people always pay for the human touch, stop worrying. Both answers are sales copy. Neither one will survive contact with your actual business plan, because neither one is about your business at all.

I want to give you a third answer, which is a method. A specific, repeatable way to score any business idea from zero to ten on AI exposure, the same method this site ran on every course in its catalog before publishing it. The score is not a mood and it is not a verdict on your worth. It is a number with bands, and each band has a different instruction attached.

The fear is not imaginary

Before we build the method, let's establish that the question deserves a serious answer, because the disruption people keep warning about is measurable in at least one place: how customers find businesses at all.

In the first four months of 2026, 68.01% of Google searches ended without a click. The search engine answered, or an AI summary answered, and nobody visited a website. In 2024 that number was 60.45%. In 2019, when SparkToro first measured it, it was 49%. The floor has been falling out from under a specific business model, the content website that lives on search traffic, and the fall is documented, not predicted.

It shows up on the other side of the same coin, too. When Google shows one of its AI Overviews above the results, the top-ranking page now gets roughly 58% fewer clicks than the same position earned in late 2023, per Ahrefs' analysis of 300,000 keywords. An entire generation of small businesses was built on the assumption that ranking well meant getting found. For anything whose customers arrive through informational search, that assumption is now worth re-checking with real numbers.

So no, I will not tell you the fear is silly. For some business models the fear is a receipt.

The fear is also badly aimed

Here is the part the panic sellers skip. In the same two years that search clicks collapsed, the physical service trades in our catalog, cleaning, lawn care, appliance repair, pressure washing, kept right on pricing and booking. The AI wave hit their businesses too: scheduling software got better, invoicing got easier, route planning stopped being a paper map. That made the operators more profitable, not less. A machine that does your paperwork is a subsidy. A machine that does your deliverable is a competitor. Those are different machines, and most people asking the question have never been taught to tell them apart.

That distinction, replace versus accelerate, is where the method starts. Then it gets sharper with four more questions: what the customer is actually paying for, where the demand comes from, who absorbs the blame when something goes wrong, and how much of the first year's work a machine could do without you.

Notice what is missing from that list. Nowhere do we ask whether AI is impressive, or whether it is getting better, or whether it will be conscious by 2030. I have no idea and neither does anyone selling you a prediction. The method only asks questions you can answer by looking at a specific business and its actual customers.

My disclosure promise

One more thing before we start, because you should know who is talking to you.

This guide doubles as the disclosure of how this catalog was curated. When we built the course library, we scored every proposed business on the method you are about to learn. Most scores came back in teachable range and the courses got built, each one with its exposure rating published in its own metadata. Three business types scored so high we removed them from the catalog, and I will name them and show you the reasoning. Four more were removed for a different reason, capital intensity, and I will name those too, because a catalog that prunes in silence invites you to assume it never prunes at all.

I would rather you see the whole record. If you finish this guide and run the method on one of our own courses and think we scored it wrong, that is the system working. The scores are arguments, not scripture.

The first distinction, replace or accelerate, is the one that does most of the work, and it is where the method begins in earnest.