Big beats small, unless small is fast enough to go where big cannot follow

There is a version of the AI story that most small business owners have not fully reckoned with yet, and I think it is worth saying plainly rather than softening it.

Imagine you run a large law firm. You have a thousand junior associates doing the kind of work that involves reading, researching, drafting, and reviewing. That work is exactly the kind of work AI is now good at. So you replace most of that junior workforce with an AI tool, keep your senior partners doing the judgment work that actually justifies their fees, and your cost base drops dramatically while your output barely changes. You have just gotten radically more efficient, and you did it by removing a layer of headcount that a small firm never had in the first place.

Now imagine you run a small law firm. You do not have a thousand juniors to replace. You might have two or three. The efficiency gain available to you from the same technology is real, but it is nowhere near the scale of what your largest competitor just achieved, because there was nothing equivalent for you to cut. The big firm just widened the gap between you, using the exact same tool you also have access to.

This is not a story about law firms specifically. It is a story about every industry with a labour-heavy cost base and a hierarchy of expensive senior people sitting above a larger layer of cheaper, more replaceable work. Accounting firms, engineering practices, consultancies, insurance, anywhere the traditional model relied on a pyramid of junior staff doing the volume work beneath a smaller number of experienced people making the calls. AI does not threaten the small operator and the large operator equally. It hands the large operator a lever that only works if you have enough scale for there to be something meaningful to pull.

I do not think this is a temporary phase that resolves itself. I think it is a structural shift, and small businesses that assume AI is simply a tool everyone gets to use equally are going to be unpleasantly surprised by how uneven the actual benefit turns out to be.

So where does that leave a small business.

The path that remains is narrower than most small businesses want to admit, but it is real. You cannot out-efficient an organisation that just removed a thousand people from its cost base, so competing on breadth or price is a losing game before it starts. One of the strongest positions left is to become the unmistakable expert on something genuinely narrow, narrow enough that no large competitor would ever bother building a process around it, and specific enough that when someone needs exactly that thing, your name is the only one that comes up. Ironically, the same technology that widened the gap at the top can be turned to your advantage at the small end too, not as a mass-market efficiency play, but as something built around exactly how you already work, which is a different use of the same tool entirely.

This is a different game to the one small businesses have traditionally played. It used to be enough to be a good generalist, competent across a reasonable range of things a local client might need, and personal service made up the difference against bigger competitors who were less responsive. That is no longer enough, because AI has made big competitors just as responsive and dramatically more efficient at the same time. The advantage small business used to hold through service alone has been partly absorbed by the same technology that widened the efficiency gap in the first place.

What has not been absorbed, and what I do not think ever will be, is deep, specific, hard-won expertise in a genuinely narrow domain, combined with the fact that you are a real person a client can actually reach. Not a generalist who happens to be small, but the recognised expert in one particular thing, who clients seek out by name because nobody else does that one thing as well. That kind of reputation cannot be built by an efficient process, because it was never a process in the first place. It is judgment and pattern recognition built over years inside one specific, narrow space, and alongside genuine relationships and a reputation clients trust, it is one of the few parts of the value chain that scale does not automatically win.

Big is going to keep getting the benefit of scale. Small wins by becoming genuinely difficult to substitute, whether that is through deep expertise in a narrow niche, relationships built over years, or judgment that only comes from specific hard-won experience, and by refusing to compete on the ground where scale always wins anyway.