This is the written version of a talk I give to chambers of commerce and SME associations. The argument, in short: the businesses that will benefit most from AI are the ones that will never hire a data scientist, and the reason is arithmetic.
The arithmetic
A large company can afford a team of twelve to save three per cent. A company of nine people cannot afford a team of one — but it can save thirty per cent, because the same person is doing sales, invoicing, support and the newsletter.
That is the whole opportunity. It is not about capability. It is about where a marginal hour goes.
Three places it shows up
Quotations and proposals. Most SMEs rebuild the same document with different numbers, forty times a month. This is the highest-return, lowest-risk place to start, because you already know what a good one looks like and you can check the output in ten seconds.
Customer replies. The same eleven questions, answered by hand, by the owner, at 11pm. Drafting replies is not glamorous and it gives back evenings.
The thing nobody writes down. Every SME has a process that lives in one person's head. AI is very good at turning a rambling explanation into a checklist — and the act of explaining it is where the value is, not the document.
What I would not do
Do not buy a platform before you have a task. Do not start with the thing that touches money or customers. Do not let a vendor define success in a way you cannot verify.
And do not assume your staff will resist. In my experience the resistance is almost always from the middle, where someone's status depends on being the person who knows.
The uncomfortable bit
The businesses that move first will not have better technology. They will have clearer processes, because you cannot automate a process you cannot describe. AI is, among other things, a very expensive way of discovering that your operations were held together by memory and goodwill.
That discovery is worth more than the automation.
