Why Every Business Owner Should Study This AI Example

In this episode of 5 Minutes in the Lower Middle Market, we explore one of the most eye-opening AI use cases I've seen in a traditional business.

A roofing company in South Florida recently ran a blind test on a $6.5 million commercial estimate. A senior estimator spent two weeks building the bid. An AI-powered system completed the same work in just 23 minutes. The difference between the two estimates? Roughly $400.

The lesson isn't that AI is replacing people. It's that AI is turning tribal knowledge into scalable operating systems.

We discuss why the biggest opportunity for lower middle market businesses may not be cost cutting, but capacity expansion, faster training, reduced key-person risk, and the ability to grow revenue without adding overhead at the same pace.

Timestamps:
0:00 Why most people still underestimate AI in traditional businesses
0:45 The $6.5 million roofing estimate experiment
1:10 Two weeks vs. 23 minutes: the shocking result
1:49 Why this changes more than productivity
2:21 Turning tribal knowledge into an operating system
3:01 How AI makes businesses more scalable
3:43 The hidden benefit: sending more proposals
4:00 Growing revenue without adding overhead
4:36 What AI adoption could look like in the lower middle market
5:00 The biggest lesson for owners and investors

Transcript

SPEAKER_00 0:00

Welcome to Five Minutes in the Lower Middle Market, where I break down the best ideas about buying, building, and owning small businesses. And today's episode is very special because it's about operating and growing this small business. And I still believe that a lot of owners, operators, and investors strongly underestimate what AI can do to a traditional business from an efficiency standpoint. I will be giving you an example from a roofing business, which story I just heard. And before hearing it, I thought AI in traditional businesses would mostly help around the edges, maybe a bit better summarization, maybe a bit faster paperwork, maybe better customer emails. But the roofing company was bidding on a very large commercial project in South Florida, 47 individual buildings, a huge set of plans and specification sheets, and a senior estimator with years of experience doing the work, the so-called old-fashioned way. The human estimator spent about two weeks building the estimate. The AI system, which was put together by a specialist running a blind test on the same project, took about 23 minutes. And on a six and a half million dollar estimate, the AI and the human were within $400 of each other. So think about again. For an senior estimator, the one job took about two weeks to put together. Now for the AI, you're just giving it all the information, all the same data, all the same materials, documents, and it will take you 23 minutes. The job itself, six and a half million. And the difference between AI and human is $400. So this is not about Claude or ChatGPT helping write a better email. That is a real workflow inside a real business on a real project with real dollars attached to it. And even the people building the system said they were pretty blown away by the result. Now, why does it matter so much for the lower middle market? Because the real opportunity here is not just cost savings, it is capacity expansion without proportional headcount growth. That roofing business did not use AI to pr to replace estimators. It used AI to make them dramatically more productive. The AI was pulling from architectural plans, similar historical jobs, materials pricing, labor pricing, and the job-specific risk factors, then routing all of that through an orchestrator that applied the business's own logic. In plain English, it was taking years of tribal knowledge that normally live in the head of one key person and turning that into a repeatable operating system. And that is where I believe at least this gets really interesting. Because once you do that, a business does not just get faster, it gets more scalable, it it gets less dependent on one person, it gets easier to train junior employees, and potentially it gets more valuable. That's what stood out to me most. The AI didn't just produce a closer answer, let's say once, it changed the economics of the business. Junior estimators could get up to speed faster, senior estimators could review work much faster, and more time could be spent with customers instead of staring at the spreadsheets. And most importantly, the company could send out more proposals. That was actually one of the biggest practical outcomes. Their customers wanted them to bid on more jobs, but without AI, that would have required hiring a much bigger team before the revenue was even one. With AI, they could respond to more opportunities without bloating overhead. That is a huge lower middle market lesson because so many businesses hit the same wall. We could grow more, but to grow more, we need to hire all this overhead first. And often that is exactly where growth stalls or growth stops. What is interesting here is that this roofing business was able to push through that ceiling. And according to the discussion, they've been able to send out more proposals, win more work, and the business is now expected to grow significantly next year. And AI in the lower middle market may not first show up as some dramatic replacement of entire teams. It may show up as one estimator doing the work of several junior employees becoming productive much faster, so they could instead of just sitting in front of the computer, they can just go and talk to customers. Owners making better decisions and making them faster. And businesses growing without the same overhead track. So if I had to pull one lesson from this episode, it would be that AI is not just a nice productivity tool for lower middle market businesses. In the right workflows, it can dramatically increase efficiency, reduce dependence on key people, and let the company grow faster without adding the same amount of overhead. And that 6.5 million roofing estimate is the kind of example that should make every owner, operator, and private equity professional pay attention. Because if AI can get within $400 of seasoned estimators answer in just 23 minutes instead of two weeks, then the efficiency upside for traditional businesses may be much bigger than most people think. That's it for today's five minutes in the low middle market. Talk to you again in the next one.