AI isn't being adopted faster for a human reason that has little to do with the technology, and I see it every time I talk with people who sit in decision-making positions at large companies. A good part of the work that thousands of people do today could already be handled in a different way, and yet almost nobody dares to lay off large numbers of employees in order to replace them with AI.
Nobody wants the mess
Laying off thousands of people creates an enormous mess, and since the credit for a decision that goes well gets spread among many people while the blame for one that goes badly lands on whoever signed it, the person deciding knows they are exposing themselves alone. If something breaks, if a process fails or if an important client leaves, someone has to explain what happened, and that someone would rather not be them. For the same reason they don't want to be responsible for whatever comes afterwards, so the most comfortable option ends up being to touch nothing and carry on as before.
On top of that, the CEO job, or any senior role in a big company, is quite political, and whoever holds it usually hopes to stay in place long enough to retire and live well. They have a few years left, five maybe, and they won't put those at risk for a change whose results will be collected by the person who comes next.
I say this without any intention of criticizing anyone, because I find it a fairly rational way to behave given that the incentives are set up exactly like that.
A familiar pattern
This has happened before. Kodak invented the digital camera in 1975 and didn't push it because it threatened the business that paid the bills, so even though it knew what was coming, it chose to stand still. Something similar is going on with AI, with the difference that what's at stake now is people rather than a product, and established companies, who know perfectly well what can be done today, prefer not to pay the political price of doing it.
Meanwhile, others are doing it
Independent entrepreneurs and freelancers, on the other hand, are bringing AI into everything they can, which makes sense when you consider that it lowers their costs and lets them scale without adding headcount. One person, or a team of two or three, now does what used to take ten, and you can see it in how fast they ship new products.
I don't know whether it's accurate to talk about geometric growth in the strict sense, but the leverage they get seems real to me. Instagram had 13 employees when Facebook bought it for a billion dollars, and that happened long before the AI we have now, so a comparable team today could be even smaller. Someone working this way has nobody to lay off and no board to explain their decisions to, and an idea, good tools and the drive to grow are enough.
What's going to happen
For all of this I think the following will happen: someone will build on their own, in a few weeks, the same product that a company with thousands, tens of thousands or hundreds of thousands of employees offers, or a better one, and will compete with it head to head. The only things setting them apart will be the brand, the branding and the personal relationships the big companies have built over many years.
That carries real weight, especially in large sales where trust and contacts tip the balance, although it won't sustain an advantage indefinitely. Once a customer compares price, speed and quality and finds equivalent products, the brand by itself starts to become expensive to maintain.
Where I land
This is what I see, and I offer it as a personal reading without any claim to predict things exactly. The technology is already available, the limitation sits in the interests of the people who can make the decision, and that limitation doesn't reach someone starting from zero. If big companies don't dare, others will do it in their place, and they will do it without asking for permission.