Something strange is happening at work. AI adoption is climbing fast, and the people meant to use it are falling behind even faster.
Roughly three-quarters of small businesses now use AI in some form. In the same breath, they admit they need training and support to get real value from it, according to the Goldman Sachs 10,000 Small Businesses survey. On paper, the revolution has arrived. Underneath runs a quieter problem: the tools change faster than the people using them can learn.
The World Economic Forum's Future of Jobs Report 2025 puts a figure on the strain. By 2030, 39% of the skills a job needs today will be reshaped or out of date. Of every hundred workers, fifty-nine will need retraining, and eleven of them will get none at all. Saadia Zahidi, the Forum's Managing Director, sets out what hangs on it: "The global economy is undergoing its most significant transformation in decades. But the future of work is not fixed. How it unfolds for workers depends on opportunities for learning, support for job transitions and backing for entrepreneurship."
That one word, opportunities, is where the smallest businesses lose out.
The ask is brutal, and the smallest carry the most
Retraining a workforce for AI sounds reasonable until you picture who has to do it. Think of the nine-to-five team at a firm that runs on relationships and judgement, not code. Asking them to become prompt engineers between client calls is not upskilling. It is a second job.
The resources are not there either. The OECD finds that the two biggest barriers to AI at smaller firms barely move: too little in-house skill, and too little budget. The very things you would need to close the gap, the expertise and the money to buy it, are exactly what a small business has least of.
It surfaces as confidence. The smaller and more human the business, the less sure it is that it can adopt AI well, and the fewer hands it has to work the problem. A two-hundred-person company can spin up an AI team. A company of twelve cannot.
Even the firms already using AI are losing ground
Here is the cruel part. Adoption does not fix it. The teams already using AI are often the most worn down by it, because the ground keeps shifting. Train everyone on a tool in spring, and by autumn the tool has changed, a better one has arrived, and the workflow you just learned is obsolete. This is not a course you finish. It is a treadmill that keeps speeding up.
It is worth saying plainly, because it usually goes unsaid: none of this is a failing of your team. The people struggling to keep up are not the slow ones. They are nearly everyone. Even the small crowd who live on the frontier, who try every new model and tool the day it lands, will admit they are only ever half-current, because another one arrives the week after. A new model, a new tool, a new way of working, every few days. If the most obsessive users cannot stay on top of it, it was never fair to expect a working team, heads down and earning a living, to.
Adopting AI is the easy part. Keeping a human workforce current is the part nobody has solved.
The answer is not more training. It is less.
It helps to remember what all this is for. LinkedIn's chief executive, Ryan Roslansky, put it plainly: "We are in the early days of a world of work that is more human than before, giving us the chance to do more fulfilling work more easily and effectively with others."
More human. That is the goal, and it is the opposite of what a retraining marathon delivers. If the price of AI is that your team spends its evenings learning tools instead of doing the work only they can do, the sums do not add up.
So turn the question around. Instead of asking how to retrain everyone fast enough, ask whether you need to at all. Rather than make engineers of your people, build the AI into the work itself, around the way the team already operates, so they get the output without the retraining. The system carries the AI. The people stay human.
There is one kind of learning worth doing, and it is the opposite of the treadmill
Chasing every new tool is the brutal kind of learning, and you should not have to do it. But there is one shift that is different, and it is worth knowing about, because it is the antidote rather than more of the same. It is the small skill of pointing AI at your own work: telling it, in your own words, what you need done, and letting it run. You do not have to follow the whole field. You only have to know how to aim it.
And you are more capable of this than you think. If you can wrestle a spreadsheet, or fight your way through the clunky software you already use every day, you can put AI to work at the scale that actually matters to your business. This is not the territory of engineers anymore. It took us about a month to learn the thing that changes everything, and it comes down to a simple idea: the moment AI can reach out into the open web and act on your behalf, the world opens up. After that, the limit is what you think to ask for, not what you were trained to do.
This is where we come in, and we mean it as a choice, not a leash. We can build the system and run it for you, so the treadmill stays ours and never becomes yours. Or we can set it up and show you how to run it yourself. Either way you come out free: never stuck, never dependent on anyone else to keep your own business moving. The aim was never to replace anyone, and never to make you need us forever. It is to lift the frantic keeping-up off your desk so your people can do the work only people can do.
The re-education gap is real, and it is widening. The businesses that come out ahead over the next few years will not be the ones that trained the hardest. They will be the ones that did not have to.