Say the word “consultant” to most business owners and you can watch them reach for the exit. I understand why. The picture in many heads is a big team, a long engagement, a very large invoice, and a deck that sits unread on a shared drive six months later. Some of that reputation was earned.
I’ve been on the buying side. I’ve watched a consulting firm come in for a small piece of work and send an invoice that felt astronomical for what we got. Now I’m on the other side, and the reaction I hear from small businesses is that consultants are out of their league or simply too expensive. Mid-market companies often tell us they’d rather hire the person. We’ve seen that happen on an engagement, and sometimes it’s the right call.
If you think a consultant is just a cost, and you can do the work internally, then do it. The real question is what you’d be buying: the time, or the thinking.
Why AI makes that question sharper
Nobody knows all of AI, and no company does either. Many organisations are fumbling their way through it. In enterprise, the push often comes from above: a private equity owner or a board asking why the business isn’t using AI, or isn’t using it well. That’s a hard position to be in. You’ve been told to move and nobody has told you where to start.
That’s where outside experience earns its place, and it’s also where AI has changed how consultants work.
The slow part has got quicker
A large share of any engagement has always been legwork: gathering data, pulling it into one place, comparing options, checking one source against another. It matters, but it isn’t where the insight comes from, and it’s the part clients have historically paid most for.
AI is very good at that part. Work that used to take days of collating can now be done in hours.
The thinking is still ours
When people hear “consultants use AI”, they picture someone pasting a question into a tool and forwarding the answer with a logo on it. Some will work that way, and I’d be wary of them too.
That isn’t the value. The value is in how you work through the material. You ask the data a question and see what comes back. You ask it to argue the opposite case. You ask what would have to be true for this option to work, where the evidence is thin, and what changes if volumes double or a key supplier leaves. Each answer tells you what to ask next.
I think of it as prompting your way to a decision. The prompts are the thinking, written down. A good consultant knows which questions are worth asking, and AI lets you ask far more of them.
We still work to the same principles as before, starting with our Operational Excellence Blueprint. AI speeds up the process. It doesn’t replace it.
What happens to your information
It’s the first question a careful client should ask, so here’s our answer. We’re extremely cautious with client data. We use Microsoft Copilot in a UK tenant, which many businesses around the world rely on. We also use Claude, on a business plan where our inputs aren’t used to train the models. If a client would rather we didn’t use Claude, that’s fine. If a client prefers us to work inside their own AI tool, we’re happy to take a licence from them and do that.
Why that reduces risk
AI isn’t always right. It often isn’t, and it delivers wrong answers with total confidence. That’s why I keep saying AI doesn’t fix broken operations, it amplifies them. Feed it messy data and you get fast, polished nonsense.
But more data points and more paths explored means fewer blind spots. Options that would once have been dropped because nobody had time to look at them properly now get a fair test, and you find the weak spots in a plan while changing it is still cheap. That only works if the data is sound and a person is judging it. More of the wrong data just makes the wrong answer more convincing.
I’m not going to put a percentage on the saving, because I haven’t measured it in a way I’d put my name to.
Small businesses shouldn’t be left out
Many small businesses can’t justify AI licences and the time to learn them yet. So where it helps, we run the work through our own tools and they get the benefit without the cost. As a new boutique consultancy, I can’t overstate how much that has helped us grow, and it has shown us what AI can do for businesses that are usually the last to get it.
A test you can use
If you’re talking to a consultant, ask how they reached their recommendation. Ask what questions they put to the data and what they tried that didn’t work. Ask what happens to your information. Someone who has done the work will have plenty to say, and someone who hasn’t will give you a summary.
I’d like to hear where you’ve seen this go well, and where it’s gone badly. Comments are open.