Over the past year, I've had an unusual vantage point. I work with B2B consulting firms trying to build or grow AI practices.
Some have been around for decades and are figuring out how AI changes an established business. Others have deep expertise and are trying to turn it into something the market will actually buy.
And I'm starting to see a pattern.
The firms getting traction aren't really entering the AI market. They're finding the intersection between something they already have unusual authority to solve and a business problem AI has materially changed.
That's a subtle distinction, but commercially, I think it's a really important one.
The trap: starting with AI
When a consulting firm decides AI is a growth opportunity, the natural instinct is to start building the AI business.
What AI services should we offer?
Should we do assessments? Training? Governance? Agents? Implementation?
Who buys AI consulting?
What should the AI section of our website say?
Then comes thought leadership, webinars, LinkedIn content, events, lead generation. And suddenly you've built an entire go-to-market motion around a market you haven't actually proven you can win.
I'm increasingly convinced that's backward.
Start with the problems you've already earned the right to solve
I've started asking consultants a different question:
Think about the last five to seven substantial engagements someone paid you to do. If today's AI capabilities had existed then, what would have been different?
Would the client's problem have changed?
Would the way you solved it have changed?
Would a bottleneck have moved somewhere else?
Would the client have needed a capability they didn't need before?
Would the economics have changed?
Would there be a new risk?
Would the outcome suddenly be achievable much faster?
That's a much more interesting place to look for an AI consulting business.
Because now you're not trying to convince the market that you're an AI expert who can solve its problems.
You're starting with a problem you already understand deeply and asking how AI has changed it.
Three commercialization traps I'm seeing
1. AI-first positioning
“We help enterprises adopt AI.”
“We help organizations become AI-native.”
“We help companies unlock the value of AI.”
The problem isn't that these statements are wrong. It's that an enormous number of consulting firms can credibly say them.
AI is the context. It isn't necessarily the reason someone hires you.
The more interesting question is: What has become newly difficult, expensive, risky or possible because of AI? And why are you unusually qualified to help with it?
2. Capability-first offers
Assessments. Training. Governance. Agents. Strategy. Implementation.
Those are capabilities.
Buyers may ultimately purchase one of them, but they're rarely waking up worried that they don't own enough assessments. They're worried about something happening in the business.
One of the fastest ways I've found to pressure-test an AI offer is simply to ask, "What business problem does this solve?"
If the answer immediately turns into a description of the service, keep digging.
3. Scaling before finding the first win
This one is particularly tempting because everything looks like a marketing problem.
We need better positioning.
We need a new website.
We need content.
We need campaigns.
Maybe...
But for smaller consulting firms entering a new market, I've become a big believer in getting extremely narrow before getting loud.
Pick the market where you have an unfair amount of credibility.
Pick the buyer experiencing a problem you understand unusually well.
Form a hypothesis about how AI has changed that problem.
Build the smallest credible offer around it.
Test the message.
Get one win.
Learn from it.
Then decide what deserves to scale.
Your old business may contain your AI business
This may be the biggest pattern I'm seeing. The consulting firms with the most interesting AI opportunities often don't need to invent an entirely new business.
They need to look harder at the business they've already built.
A firm with years of experience in regulated industries knows things about risk, governance and accountability that a brand-new AI consultancy doesn't.
A firm that has spent a decade fixing delivery problems knows what happens when organizations suddenly increase the speed at which work can be generated.
A firm deeply embedded in a particular function understands workflows, incentives, decision rights and organizational realities that someone arriving with an AI toolkit won't.
So if I owned a consulting firm trying to figure out its AI growth strategy right now, I wouldn't start with, “What AI services should we sell?”
I'd put the last several years of client work on the table and ask "What problems have we already earned the right to solve and how has AI changed them enough that our clients need our help again?"
I suspect there's a much better business hiding inside that question.