The economists say your kind of work wins. Your clients just won't know you exist.

You run a salon, a repair business, a small clinic, a kitchen, a property service. You are good at it, and most weeks you are too busy to think about "marketing." Work has always come in the same way: a friend passes your name along, someone sees your sign, someone searches and calls. You fix the boiler, cut the hair, plate the dish, show the flat. The next job follows the last one.

Something quiet changed in the last year. When a neighbour needs an electrician, a physiotherapist, or a place for a birthday dinner, more of them now open ChatGPT or a similar assistant and just ask it: who should I use for this. The assistant answers in a few lines. Two or three names, sometimes a reason for each. If your business is not in that short answer, you were never in the running. And unlike a missed call, you will never know it happened.

What the model actually says

This is not a hunch, and it is not a sales pitch. Economists at Anthropic, working with researchers from several universities, built a model of how advanced AI could reshape the US economy over the coming years, and published the scenarios openly. In their picture the economy grows and average pay rises, but the gains do not spread evenly. The money moves toward hands-on, in-person work, and away from work that happens on a screen.

The study's authors, Anthropic's Economics team, put the losing side plainly:

Most knowledge workers face either lower wages or unemployment.

The other side of that sentence is the part that matters to you. As desk work gets automated, the people who build, fix, treat, cook and serve in person become relatively more valuable, not less.

Why the trades and in-person services come out fine

The reason is almost mechanical, and it is worth understanding once. AI is fast and cheap at anything that lives on a screen: drafting, summarising, analysing, writing code. It is useless at rewiring a fuse box, colouring hair, resetting a shoulder, running a busy service on a Friday night, or walking a buyer through a flat and reading their face. Those things still need a real person in a real place.

When the screen work gets cheaper, two things happen at once. The businesses and households that save money on it have more to spend, and a larger share of what they still want to buy is precisely the work a machine cannot do. So demand for hands-on services holds up, and in the stronger scenarios it grows.

The scale depends entirely on how fast the technology moves, and the model is honest about the spread. It runs three futures. In the mildest, US output is about 1.6 percent higher by 2030. In the middle one, 8.3 percent. In the most extreme, 32.4 percent. In that extreme case the share of national income going to workers rather than to the owners of capital slips from roughly 60 cents on every dollar toward 45. But read where that squeeze lands: on the replaceable desk jobs, not on the plumber, the nurse, or the baker.

One honest caveat, because it buys the rest its credibility. This is a single model of the US economy, an early first version, and its authors deliberately leave out robotics, government policy, and the shocks that always arrive. It is a well-argued direction of travel, not a promise about your street. Treat it that way.

The catch: more demand does not help if you are invisible

Here is the turn that should hold your attention. More demand for your kind of work is only good news if the demand can find you. And the front door is moving.

For twenty years the front door was a search page: ten blue links, and being near the top was enough to get a look. Increasingly the front door is a single AI answer that names two or three businesses and then stops. Ten links gave you a chance to be noticed. One answer that leaves you out gives you nothing, and it does it silently.

The assistant builds that answer from whatever it can read about you across the open web. Your own site. Your map and directory listings. Your reviews. The plain facts of what you do, who you do it for, and where. If that trail is thin, contradictory, or a year out of date, the assistant does the sensible thing and recommends the business whose trail is clear. You are not losing to a better electrician down the road. You are losing to a better-documented one.

That is the real shift behind all the noise about AI. For a local, hands-on business, the threat was never that a machine takes your job. The threat is that the machine quietly stops mentioning you.

What to do this month

None of this needs a budget, an agency, or a free weekend you do not have. Five things, and you can start today:

1. Say plainly what you do and where. Somewhere a machine can quote you, in one clear line: the service, the town or neighbourhood, who it is for. "Boiler repair and servicing in Girona for homes and small cafes" works. "Your comfort is our passion" does not.
2. Make your basic facts agree with each other. Name, address, phone, opening hours, service area, identical on your website, your map listing, and every directory you appear in. Assistants trust facts that match across sources and quietly distrust ones that contradict.
3. Publish a little, in plain language. A handful of short, honest pages about the questions clients actually ask you: what a job costs, how long it takes, what usually goes wrong. This is the material an assistant reads to decide you genuinely know your trade.
4. Collect recent reviews, and make leaving one effortless. Not a campaign. One link, sent at the moment a client is happy. Fresh, specific reviews are a strong signal to people and to the machine reading on their behalf.
5. Ask the assistant about yourself. Open ChatGPT, or another one, and ask it as a customer would: who should I hire for this service in this town. If you are not named, or the details are wrong, you have just found your real starting line.

Where Laspi Pro fits

That list is the entire job, and you can do all of it by hand. Laspi Pro exists to make it take minutes instead of evenings. You talk, in your own words, about what is new in the business. It keeps your facts consistent everywhere, writes those plain-language pages and posts in your voice rather than in generic ad-speak, and shows you what the AI assistants currently say about you when a customer asks. It will not put you in the answer by magic, and it does not pretend to. It makes sure the trail about your work is clear, current, and easy for both people and machines to read, which is the part that was quietly costing you jobs.

The economists' bet is that the next few years reward exactly the work you already do well. Whether that reward reaches you comes down to something smaller and far more fixable: whether, the moment a client asks the machine, your name is there to be found.