Own a Niche, Own a Place, or Own a Stack

October 4, 2026

Read Time: 8 minutes

Something is changing here.

This newsletter has been one thing for a while now. How to build an AI consulting business, and how to get clients for it. That is not going anywhere. If it is why you subscribed, nothing you came for is leaving.

But it has become too narrow for what is actually happening out there.

The same questions keep arriving from people who are not consultants. Owners trying to work out where to start. People who want to build something in AI and are not convinced a consultancy is it. Operators who have been handed the AI problem internally along with no budget to solve it.

So Get It Dunn is widening. Still AI consulting at the centre. Now also AI opportunity more broadly, which means starting an AI business and adopting AI inside one you already run.

A few things coming:

• A shorter news and update edition, probably twice a week

• Workshops on getting genuinely good at Claude, Viktor etc…

• How to start a local AI agency and stack recurring revenue.

• How to build and ship your own software to simplify your business/workflows.

What does not change is the deal. One insight, and something you can actually do with it. If an issue ever turns into a think piece, hit reply and tell me off.

Now let’s get into this week’s newsletter.

I know a man who built a consulting firm doing $32 million a year with $16 million of EBITDA.

Not an AI business. He consulted for organisations that used Salesforce and had sales teams. That was it. That was the whole thing.

Two constraints, both of them narrowing. He only worked with sales teams, and he only worked on Salesforce. Everybody else in the market was free to take any client with any system, and he was turning down most of them on principle.

He sold it for $162 million.

I think about that firm a lot, because it is the clearest example I know of something most consultants never consider. There are three ways to narrow, not two, and almost everybody only knows about one of them.

The Third Way to Narrow

Back in August I wrote about owning a place rather than a vertical. A local moat, in person, referral networks, local press. That was the second option sitting next to the one everybody already knows, which is owning a niche.

There is a third, and it is the stack.

Not the industry your client is in. Not the city they are in. The software they run their business on.

And you can take any combination of the three. Own a niche. Own a place. Own a stack. Own a niche and a stack, which is exactly what the Salesforce man did and exactly why that business was worth what it was worth. Own all three if you want.

They stack on top of each other rather than competing, and each one makes you harder to compare against the next person.

You Do Not Need the Experience First

Worth clearing this up, because it is the reason most people skip the idea.

You do not have to have spent eight years inside an ecosystem to claim it. If you already have that background then you are sitting on positioning you have never used, and you should go and use it. But choosing one from a standing start is an entirely legitimate move.

Pick one. Go deep. That is the whole exercise.

Microsoft is where I see this show up most often with the consultants I train, and for a straightforward reason. An enormous number of traditional businesses are on the Microsoft suite, they started there, and they are not moving. Ever. So if you want to work with those businesses, you have to learn to operate inside that environment.

I will be blunt about Copilot. It is not good, and the suite is behind where everyone else is right now. That is not an argument against the stack, it is the argument for it. Somebody has to be the person who makes AI work properly for a manufacturing business that is never leaving Microsoft, and right now that person does not really exist.

Google I think is a genuinely good bet. It is behind in some respects and it is not as capable as Anthropic or OpenAI at certain things. But Gemini is excellent, image generation is the best of the lot, and Google has a long track record of being extremely competitive once it decides to be. I would not want to be short Google over five years.

Anthropic or OpenAI as a declared specialisation works too. Same mechanic, different ecosystem.

All four of those are available. None of them are crowded.

What Changes on Monday

Less than you think, which is why this is worth doing.

Your actual delivery barely changes. You are still mapping workflows, still finding the bottlenecks, still building. What changes is who you say it to and what you call yourself.

So instead of this:

"AI consultant helping businesses improve efficiency."

You get this:

"Microsoft AI consultant for manufacturing companies."

Nine words, two constraints, and it is immediately obvious whether it is for you. A manufacturing operations director reading that does not have to work out whether you understand his environment, because you have told him.

Then your targeting follows the line. You are no longer looking for businesses that might want AI help. You are looking for manufacturing companies running the Microsoft suite, which is a list you can actually build.

That is the whole change. The headline and the list.

The Ecosystem Is the Lead Source

Here is the part that makes this better than sector niching alone.

A vertical does not come with infrastructure. A stack does.

Partner programmes. Marketplaces. User groups. Events. Established implementation partners who are good at their platform and have no idea what to do about AI. All of it already exists, all of it is organised, and all of it is full of people who have self-identified as being on that stack.

In my experience the ecosystem becomes a lead source in its own right, and a far more reliable one than cold outreach. You are not interrupting strangers. You are turning up somewhere people already gather around the thing you have decided to specialise in.

Nobody runs a user group for mid-sized businesses in general.

The Obvious Objection: Doesn't This Shrink Your Market?

Yes. That is what narrowing means and there is no way around it.

And the first few months genuinely feel worse. You are turning down work you could have taken. The pipeline looks thinner. Every instinct tells you that you have just made the job harder for no reason, and that feeling is real rather than imagined.

Here is the counterweight.

Being known for one thing means you eventually charge significantly more for the same work. Not marginally more. Significantly. Because you are no longer being compared against three generalists on price, you are the person who does this specific thing, and there is nobody obvious to put next to you.

And people come to you rather than the other way round. That compounds quietly and then all at once.

A slower start quite often produces faster growth. I have watched it happen enough times to be comfortable saying it.

Honestly, I do not think there is much of a downside here beyond the discomfort of the first six months. You can construct an argument against any form of specialisation if you want one. I would rather be the Microsoft person for manufacturing than the fourth AI consultant on somebody's shortlist.

The Real Point

There are three ways to be specific, and most people only know about one.

The industry. The place. The software they run on.

Pick a stack, go deep, rewrite one line of your positioning, and go and find where that ecosystem already meets.

One man’s firm did $32 million a year on sales teams who used Salesforce. He did not need a bigger market. He needed a smaller one.

See you next week,

– Andrew