The Dealmaker’s Guide to Stress-Testing AI Value

If you ask private equity firms or their portfolio companies whether they’re using AI, the vast majority will say yes, but many still aren’t seeing significant value from the technology. GrowFast Advisory’s Stephen Murray is back on the podcast to share how to identify real AI-generated value, retain and nurture that value, and track the AI metric every leader should watch.

Stephen Murray: Thanks, Carolyn. We maximize enterprise value for an exit in a post-AI world. From that perspective, every deal already has three diligence workstreams: quality of earnings, legal, and commercial.

We believe there is a fourth workstream that has not received enough focus: Where is value actually captured, and is it transferable? We would say that is particularly important today in the world of AI. I’ve personally sat in three chairs in a deal.

I’ve been in the seller’s chair across from private equity, the buyer’s chair across from Fortune companies, and, perhaps most uniquely, the chair responsible for executing the deal. That pulled back the curtain and revealed many issues and places where value can be lost. We focus on that problem.

We consider the individuals who created the value, the work systems, and, particularly today, the way AI and technology are changing the workforce so quickly. We examine the issues that change creates for a company.

SM: Easy: “Under Pressure,” whether it’s the Queen or David Bowie version. The race to get ahead in AI is constant every day. But if you’ve sat in a management meeting and discussed an upcoming deal, the question becomes: Who actually runs the company? Outside of the CEO, you’ll often find the room gets quiet. That bassline from “Under Pressure” starts building in your head. I always come back to that song.

SM: About a year ago, many people focused on AI adoption rates and assumed that the more a company used AI, the more effective it had to be. The conversation has since shifted to whether the company gets value from what it has invested. I’ll slow down here and share a statistic that I found fascinating: McKinsey’s latest number indicates that 88% of companies are using AI, but only about 6% are getting real financial value from it.

If you pause and consider that, it’s a massive gap. A company can be using AI, but if it didn’t implement it correctly, or lacked the diagnostics, foresight, and structure, it might as well not have implemented it. It may simply have been chasing the shiny object.

The companies starting to see major value are the ones that carefully considered what they needed and how AI would affect different parts of the business. It comes down to questions such as: Who owns it? Which work system does it improve?

What are competitors doing similarly or differently? The companies creating value are those that examined what sits behind the technology before implementing it.

SM: Let me tell a short story about identifying value. Often, it comes down to observation as much as questions. I recently spoke with a very sharp CFO and asked what it costs to get a quote out the door today compared with two years ago.

There was a long pause. He didn’t know. Then I asked what the company had stopped doing since it bought AI.

There was an even longer pause. That question gets to the relationship between old and new processes. If you implement AI, it should eliminate an inefficiency or replace something you were already doing.

If it doesn’t, you’re simply paying for two systems and adding another subscription. In a conversation lasting less than 90 seconds, you can begin to determine whether revenue per employee has increased, whether cycle time is improving cash flow, and whether EBITDA is changing. Look at the management team.

Do leaders understand the deeper processes within their own organizations where AI has been implemented? These are basic questions. If they don’t know the answers, the business could be discounted during a deal because they can’t identify the people who built the system or explain what the system replaced.

SM: That is the question. Every AI workflow that matters has a person attached to it. Who owns it?

I don’t mean the executive sponsor. I mean: Who built it? Who fixes it at 9 p.

m.? Who knows why it was configured the way it was? Imagine buying a used car in a parking lot.

It looks perfect and has been kept up to date. You kick the tires, and it passes every test. But as soon as you want to drive it off the lot, you discover the person who built it has left.

Now you’re holding something that may not mean very much. Diligence is easier than people think because it starts with knowing the people, understanding the audience, and asking the right questions. Those questions lead to answers that help you uncover what traditional diligence may miss.

SM: The consequences can be significant. In one deal, we bought an advanced manufacturing organization and invested in robotic arms and AI. We determined where to deploy the technology, but we failed to account for the fact that engineers, rather than typical operators, would run it.

Engineers have different compensation structures, benefits, training needs, recruiting costs, and other employment requirements. The technology may also replace employees who are no longer needed. If those considerations aren’t addressed from the beginning, they can create substantial unknown and hidden costs and completely erode the value you expected to gain.

It can have a significant effect on fixed and variable costs and on EBITDA.

SM: Owner coverage. For every AI workflow that affects revenue, costs, or customers, how many people can run it? Do you have a layer of people below the owner who know how to operate the technology? It sounds simple, but you should begin tracking that today in case key people leave. Go watch the work. Don’t rely only on the CEO’s explanation. If the value story exists only in a deck filled with flashy images of the future or robots, I would put my wallet away. That doesn’t get beneath the surface of what will actually transfer the value, which is people doing the work.

SM: I would say it’s the metric leaders already track: EBITDA. Leaders should also understand owner coverage. AI could help an owner step away, or it could make that person even more important to the business. What is changing in EBITDA? What is changing about the owner’s importance to the business? Several factors contribute to those answers, so leaders need to watch them closely. How much did the company spend, and did the investment affect other critical measures? Those questions begin to pull back the curtain on the value that may be present.

SM: Go watch the work. Very simply, don’t overcomplicate it or assume the technology is too difficult to understand. Look at who is running the plants and distribution centers. What is said at the conference table isn’t always what is seen on the shop floor. Make sure you understand the reality. It sounds easy because it is, if you take the time to observe it. I’ve done that many times in my career. Break away and speak with the operators. You’ll learn much more about what’s working, what isn’t, and the issues employees are facing. Have someone on your diligence team go do that.

This transcript was prepared by a transcription service. This version may not be in its final form and may be updated.