A CEO Said Data Scientists. The Room Shifted.

On a panel at the ACEC Fall Conference in Phoenix, the CEO of a large engineering firm said something I keep coming back to. More or less: if they could afford it, they'd be more interested in hiring data scientists than more principal engineers.

Nobody gasped. Nobody walked out. But you could feel the room shift.

Think about who was sitting there. Firm leaders, many of whom came up as exactly the person that CEO had just moved down the hiring list. In this profession, the principal engineer is the firm's reputation in a pair of boots. And one of their own had just said, out loud, that the next hire they really wanted might be somebody else.

I don't think that CEO was saying engineers matter less. I think they were saying what most of the questions in Phoenix were circling. A principal engineer is a bet on what one person knows. A data scientist is a bet on what the whole firm knows, if the firm has done the work to make it usable.

Most firms haven't done that work. Your people can get faster while your firm stays right where it was.

What did engineering firm leaders ask about AI in Phoenix?

AI was one of the top themes of the conference. I heard these in our session's Q&A, at other presentations, at networking events and talking with other attendees. Paraphrased from my notes:

• How do we make this work in our firm?

• What's the right approach?

• How will this enable us to compete differently?

• How does AI help us solve the talent gap?

• What does this mean for new engineers, architects and construction professionals?

• Do we just deploy an LLM chatbot and let individuals figure it out?

• How does the firm benefit?

Those are honest questions, and the talent one has numbers behind it. The ACEC Research Institute reports about 184,000 engineers retired or left the United States profession in 2022, against 166,000 new graduates coming in. In its Q1 2026 sentiment study, more than half of member firms said they're investing in dedicated AI-focused talent.

That's a wider group than data scientists, and it's firms describing themselves. Still, that CEO isn't the only one looking at that kind of hire.

The question I'd put on top is the last one. How does the firm benefit? Answer that and the rest of the list gets easier. Skip it and you're buying tools for individuals and hoping the firm catches up.

What wasn't discussed enough?

The work.

It takes focus. It takes real hours from people who already have a full plate. It isn't a side project for whoever's most excited about AI.

What engineering firm leaders aren't discussing enough is the work underneath AI: understanding and mapping workflows, organizing data, and the real investment behind both. That work is what stands between a firm and organizational intelligence.

There's something in it for your people too. Nobody became an engineer to be a biological router of data. But that's a lot of the job right now: finding the file, tracking down who knows, carrying the answer from one system to the next. Do that work and they get to be better at what they want to be doing.

The chatbot question I can answer flat. No. Chatbots handed out one person at a time create more silos, because what each one learns stays with that person.

We've seen what 50 of those look like. Everybody's faster. Nobody's connected.

Go back to the question I'd put on top. How does the firm benefit from fifty private assistants? Not much.

The move worth making runs the other way, across the firm. Connect your data sources and your workflows, and you activate what's actually yours: your firm's intelligence and the unique way you do business. That took years to build, and it's sitting in your files and your people's heads.

That's your defensible moat. A competitor can buy the same chatbot. They can't buy that.

Where the firms I talked to are stuck

Here's my read after all those conversations. Firms are still struggling to accept that this technology is different and asks for a different approach.

Small and mid-sized firms are hunting for the right starting point and a partner who'll go on the journey with them. Until they find one, single-purpose tools feel safer.

Large firms are throwing money at it, and in some cases people. What I didn't hear from them was a clear view that the future is organizational intelligence, not just faster gears automated by individual or team agents.

What both groups are missing is this. AI itself is what makes this work doable now. Mapping workflows and organizing what a firm knows used to be out of reach or cost prohibitive for most firms. It isn't anymore.

So the biggest hurdles I saw weren't technical. They were willingness, how far a leader can see given what they understand about the tech, and whether anyone will just start.

Where should an engineering firm start with AI?

Just start. Pick one workflow, define one set of data, and begin building your firm's way of doing business into a tool. Mapping the workflow and defining its sources of truth can happen in parallel, so neither has to wait for the other.

That's the part people don't expect. You don't need the perfect map before you touch the data, and you don't need clean data before you describe how the work actually runs. Each one teaches you about the other, which is why we work this way at Universal Mind.

One workflow is a bigger ambition than it sounds. It's how a firm finds out what this takes.

This work is worth the investment, and it won't show its real value in 30 days, or 90. That delayed ROI isn't because of the technology. It's because your firm has a history, an approach, processes and expertise in how it does its work, and all of that deserves the respect of being fully mapped.

Then your experts validate it, including the principal engineers that CEO had moved down the hiring list. That step is how you bring them into the process.

I don't know whether that CEO will ever hire those data scientists. They did say if.

And I don't think the answer is fewer principal engineers. This profession can't spare them. What it can't keep doing is letting what they know walk out the door with them.

But here's the question I'd take back to the office. Forget who you'd hire next. If your best principal engineer retired tomorrow, what would your firm still know?

---

Where to go next: an AI Workstream is the first step. It defines one workflow end to end in three to five weeks and ends with a scoped recommendation for what to build.

Holly Buck and I presented "Ignite Innovation: Moving from AI Skepticism to Organizational Intelligence" at the ACEC Fall Conference in Phoenix on October 26.