Most writing about AI and leadership is aimed at large companies: governance frameworks, model monitoring, transformation programmes.
For a small business, the questions are more immediate. What should I let it do? What should I check? What do I tell my team? And what does it change about what’s actually valuable?
It’s a leadership question, not a technical one
Using AI tools is easy. Deciding what to use them for is not, and that decision is a leadership one.
Every time you hand something to an AI tool, you’re making the same judgement you’d make delegating to a person: is this something that can be done well without my context, and how closely do I need to check the result?
That makes the most useful mental model a familiar one. Treat AI like a very fast, very capable new assistant who has never met your clients and occasionally makes things up with complete confidence.
You’d give that person real work. You’d also read everything they produced before it went out.
What it’s genuinely good at
First drafts. A blank page is slow. A rough draft to react to is fast. AI is excellent at getting you from nothing to something.
Summarising. Long documents, meeting notes, a thread of emails you need to catch up on.
Reorganising what you already have. Turning your notes into a structure, your bullet points into a proposal, your rough process into a checklist.
Routine repetition. The same kind of email, the same kind of document, the tenth variation of something you’ve done many times.
For a small business, that can free up a remarkable number of hours — which is why I consolidated my own tools onto a platform with it built in.
Where it needs you
Anything factual. AI tools produce plausible statistics, citations and quotations that don’t exist. They present invented facts in exactly the same confident tone as real ones, and there’s no signal to tell you which is which.
This isn’t a small caveat. Misattributed quotes and unsourced statistics are already everywhere online, and AI has made them far cheaper to produce. If it’s a number, a name or a quotation, check it against the original source before it goes anywhere with your name on it.
Your voice. AI-generated text has a recognisable sameness — confident, generic, faintly corporate. People increasingly notice it, and it reads as not having bothered.
Decisions about people. Hiring, firing, performance, difficult conversations. These need your judgement and your accountability, and delegating them to software is both unwise and, in some cases, legally risky.
Anything where being wrong is expensive. Contracts, pricing, financial and legal matters. Use it to prepare, never to decide.
Volume just became cheap
This is the most important change, and it’s easy to miss.
For years, producing a lot of content, a lot of proposals, a lot of anything, took effort — so volume was a signal of seriousness. AI has made volume nearly free.
When something becomes free, it stops being valuable. The things that become more valuable are the things AI can’t supply on its own: specific experience, verified facts, genuine judgement, and a real point of view.
So the owners who benefit most from AI won’t be the ones who use it to produce more. They’ll be the ones who use it to clear the routine work away, and spend the time they’ve saved on the parts that only they can do — the story from their own business, the client relationship, the decision that requires knowing the situation.
An AI tool can write a thousand-word post on pricing in seconds. It can’t tell your readers what happened the first time you raised your prices, because it wasn’t there.
Talk to your team about it
If you employ people, some of them are quietly worried about being replaced. Silence makes that worse.
Be honest about what you’re doing and why. Which tasks will AI take on? What does that free people to do? What won’t change?
The most useful framing, and usually the most accurate one for a small business, is that AI removes the tedious parts of people’s jobs so they can do more of the valuable parts. Say it if it’s true. And if some roles genuinely will change, tell people early enough to prepare.
Then help them use it well. The team that’s taught to use AI thoughtfully, with judgement and checking, will get far more from it than one left to figure it out alone.
A few ground rules worth setting
- Nothing factual goes out without a human checking the source.
- Client confidential information doesn’t go into tools you haven’t checked the terms of.
- Anything sent under your name gets read by a person first.
- Decisions about people are made by people.
Keep learning
One thing I’ve found to be true: the more you learn, the more you realise how much there still is to learn. That’s never been more obvious than it is with AI.
I don’t believe it’s going to take all the jobs. It’s a tool, and like any tool, what it does depends entirely on how you use it and how well you leverage it.
But we’re in a world where speed increasingly wins — the business that responds first, produces first and adapts first tends to get the work. So it’s worth learning how to use these tools, and learning to use them well and appropriately. Not to replace your judgement, but to free up time for it.
The owners who struggle won’t be the ones AI replaced. They’ll be the ones who decided not to learn.
The short version
Use AI for the work that doesn’t need you, so you have more time for the work that does. Check everything factual. Keep your voice and your decisions your own.
The technology will keep changing faster than any article can keep up with. The leadership question underneath it won’t: what should I delegate, and how closely should I check?
