The short version. Everyone’s selling AI efficiency. Nobody asks the bit that comes next. If it makes you ten times faster, are you ten times better off? Depends who you are. Efficiency isn’t a benefit. It’s a transfer. Right now it pools in very few hands. That changes. The win worth waiting for is the one that reaches everybody, not just some.
Efficiency isn’t a benefit, it’s a transfer
Everyone’s selling AI efficiency. Faster, cheaper, more done by lunch. Nobody asks the bit that comes next.
If AI makes you ten times faster, are you ten times better off?
Depends who you are. That’s the whole article.
Efficiency isn’t a benefit. It’s a transfer. The speed goes somewhere. It goes to whoever owns the work, not whoever sped up.
Who’s holding the bowl
Ten times faster is a prize. The only question is who’s holding the bowl when it gets handed out.
Run everyone through it.
The big corporate measures the efficiency and sees almost nothing. The studies keep stacking up, all landing in the same spot. Biggest one, MIT, last year. Companies tipped thirty, forty billion dollars into AI. Ninety five percent saw no return they could measure. Not a small return. None. The money went in at the top and the benefit leaked out the sides, to people quietly using it to make their own day easier. Nowhere a balance sheet thinks to look.
The staff member makes the efficiency and watches it float up past them. Ten times faster, paid the same. So the gain goes to the business as margin, or to the customer as a cheaper price, or it just turns into more work by Friday. They did the hard part. Someone else banked it. And study after study says the same from their side. Most employees already using their own AI at work, plenty paying for it themselves, plenty keeping quiet about it. People will fund their own efficiency. They just don’t get to keep it.
The only one who keeps the whole ten times is the one who owns what they make. The freelancer. The owner. The person at home building their own thing. Faster, and it’s all theirs.
So AI efficiency was never about speed. It’s about ownership. Speed’s just the thing that gets carved up.
I’m the example, and not the flattering kind
Tidy place to stop. Also a lie, and you’d smell it.
Owning the work decides who keeps the gain. It says nothing about who can make it in the first place. That’s the harder one. And here I’m the example, not the flattering kind.
AI has changed my life. My work, my business, what I can put in front of a client. Worth more to me than I can properly explain. And I’m a unicorn. I don’t know many people who use it the way I do. I can’t even get my own family to. People I love, every reason to want what it’s given me, sitting right across the dinner table. Won’t transfer.
Sit with that. The thing that changed my life won’t cross my own kitchen. So when a big company sends round a Monday email and a site licence and a slide that says people first, I want to know what they think they’re handing over. If it won’t cross a dinner table, a login won’t do it.
Watched an international expert in this lose his rag at a chat the other day. Wrong expectations, didn’t quite know how to ask, rattled by a tool that’s changing how he has to work. Ended up swearing into the box. An expert. Not a different species to the rest of us. He’s me, a couple of years back, minus the circumstance.
Using it and driving it are different things
So the honest answer to who AI benefits is smaller and more awkward than anyone selling it lets on.
Right now AI efficiency isn’t a benefit. It’s a transfer. Who keeps it comes down to who owns the work and who’s put in the hours to actually drive the thing. The overlap is tiny. Almost everyone’s standing outside it.
Most people are using AI. Almost nobody’s driving it. Different things. The gap between them is years of daily use, and nobody can hand you that, mandate it, or shortcut it.
The win worth waiting for
So how do we get it back to the people? The corporate, the department, the person at work, the person at home?
Time. Experience. That’s the whole answer, and it’s an annoying one, because you can’t roll it out.
Which means nobody’s arrived yet. Not the corporates with their rollouts. Not the experts posting threads. Not me. We’re all still working out where the value actually lands, and anyone who tells you they’ve cracked it is selling you a finish line that isn’t there. Don’t believe the posts. Don’t believe the experts. Don’t believe the cynics either, the ones who’ve decided it’s all a con. Same impatience, different coat.
I think it gets there. I just won’t pretend we’re there.
And here’s the part worth the wait. Right now the efficiency gets grabbed, because we’re early and only a few can drive it. As the skill spreads, the prize isn’t one side finally winning the split. It’s that it stops being a split. The business keeps its margin, the staff member gets their time back, the customer gets the better service. Efficiency that reaches everybody. Not a thinner slice to fight over. A bigger pie. That’s the only people first worth the words. And it’s worth waiting for.
I’d rather be in that, building toward it, still learning, than sell you a shortcut. Wherever you sit, running the business, doing the work, or teaching yourself at the kitchen table, the question’s the same. Reaching for the right thing, or just the newest one? Good conversation to have.
Garth Shoebridge is a digital problem-solver with 25 years of experience helping South African businesses think clearly before they build. If any of this is the question on your desk right now, start with a conversation.
P.S. There’s another business I’m involved in chewing on the harder problem under all this. Not how to use AI. How to put it into a business in a controlled way, without it going feral. And I’ll be honest about the tension at the heart of it, the same one this piece keeps circling. It could make a person ten times more efficient and still do nothing for the staff member. The optimist in me bets it buys them time for better work, and a bit more enjoyment of the job. Don’t know yet. Still finding out.