Try this experiment. Give an artificial intelligence system ten years of invoices and logged hours, and ask it which client is actually costing you money. What if the answer turned out to be the one you’re most attached to — the one who sends you a panettone every Christmas?
To ask this question, you don’t need to know what a prompt is.
You just need someone who can ask it for you, inside your own company.
In our industry, some people use a rather dismissive term for them: ignorant companies — the ones that don’t use artificial intelligence and perhaps don’t even want to.
We call them ideal clients.
They do their job, they’re not seduced by flashy demos, and they ask just one question: what changes starting Monday?
And they’re right to be wary.
According to MIT, 95 percent of AI projects launched by companies have failed to produce any measurable financial results.
Not because the technology doesn’t work, but because it was bought like a January gym membership — full of good intentions, then left to gather dust. The companies that succeeded started with a specific, often mundane problem, and got proper help along the way.
Those who went it alone failed far more often.
Even the biggest players have figured this out.
In May, the company behind ChatGPT launched a venture worth over four billion dollars that does just one thing: it sends engineers into clients’ offices and doesn’t leave until the AI actually works.
In California, they call it a “forward deployed engineer.”
In San Marino, it’s called “I’ll come by tomorrow morning.”
Meanwhile, artificial intelligence has already slipped into your company through the back door.
MIT also reports that in more than nine out of ten companies, employees are already using AI tools through their personal accounts, while only four out of ten businesses have purchased them officially.
The young salesperson down the hall is already using it more than you are — and to save time, they’re pasting your clients’ information straight into it.
Starting Monday, though, things can change.
The client who messages you on WhatsApp at eleven at night gets the right answer before they’ve even put the phone down.
They’re answered by an assistant that identifies itself as such and leaves the important decisions to you.
The quote doesn’t need to be built from scratch, because the system finds a similar one from three years ago on its own. And you no longer need to open your inbox: an assistant summarizes it for you while you wait for your coffee.
None of this requires a lab in California.
In a country of thirty-four thousand people, a supplier who sends a link and disappears doesn’t last a season: you run into your clients at the bar, and if something’s wrong, you know it before dinner. This is how we work here — including for companies from Italy and further afield that, for confidentiality reasons, we can’t name.
The first step is a single question: where are you losing the most time, or the most money?
Then we come to your office.
And if the answer is that you don’t actually need artificial intelligence, we’ll tell you so.
In California, it costs four billion dollars.
Here, it starts with a coffee.
Produced by the Strategic.sm Observatory, a San Marino-based company that brings artificial intelligence into businesses.
