The news came out quietly, the way these things usually do. McKinsey is cutting 3,000 to 4,000 positions in 2026—roughly 10% of its global workforce, the largest reduction since 2008. Bain, BCG and Deloitte are doing variations of the same thing. Slower hiring. Headcount reductions. A pulling back that the firms themselves describe, carefully, as a response to AI productivity gains.
Read that again. The world’s most prestigious advisory firms are cutting people because AI is doing the work those people used to do.
That is not a technology story. That is an operating model story. And if you lead a commercial construction firm, it is your story too — whether you see it yet or not.
The Work That Was Cut
McKinsey didn’t eliminate partners. They eliminated the analytical layer underneath them—the associates and analysts who spent weeks synthesizing research, building models, producing deliverables that clients paid for at premium rates. That work turned out to be compressible. Not because the people weren’t capable.
Because the work itself—pattern recognition, document synthesis, comparative analysis, structured reporting—is exactly what artificial intelligence (AI) does well and fast.
Now look at your own organization.
Who tracks RFIs and summarizes status for the project executive? Who pulls together the weekly schedule report? Who runs the first pass on subcontractor bids? Who writes the owner update? Who coordinates the submittal log across three trades?
That work is not the same as McKinsey analyst work. But the underlying structure is identical. Information processing. Pattern recognition. Synthesis and reporting. Work that requires training and intelligence—and that AI is now compressing on a timeline most construction leaders haven’t fully absorbed.
This is not a prediction. It is already happening. Firms using AI-assisted estimating are running comparables in hours that used to take days. Document review that required a coordinator’s full afternoon now takes minutes. RFI response drafts that sat in someone’s queue are being generated in real time.
The compression is real. The question is whether your operating model is designed for what comes after—or still built for what worked before.
What the Compression Exposes
Here is where most firms get stuck. They see the compression coming, buy a platform, run a pilot, and call it an AI strategy. The tools do what they were built to do. And then not much changes—because the tools were layered on top of an operating model that was never redesigned to use them.
The technology is not the problem. The sequence is.
An AI strategy matters. You need one. But an AI strategy built on top of a fragmented operating model does not create leverage. It accelerates the existing dysfunction. Faster reporting on a project with unclear decision rights is still a project with unclear decision rights. Better data flowing to people without authority to act on it is still a bottleneck—just a better-informed one.
McKinsey learned this at scale. They built a business model on billable analytical hours, and then the thing that made those hours billable got automated. What survived was senior judgment—the partners who could walk into a boardroom and tell a CEO something they didn’t already know.
The analytical scaffolding underneath that judgment turned out to be the exposure, not the asset. Your firm has the same anatomy. The question is whether you know where your exposure sits.
AI Needs People. Specifically, It Needs Judgment
There is a version of this conversation that treats AI as a headcount reduction strategy. That framing will cost you.
AI does not manage owner relationships. It does not make the call on a subcontractor who is three days behind and won’t say so. It does not read the room when the GC and the owner’s rep are heading toward a dispute that will cost six weeks if it isn’t defused today.
It does not carry the pattern recognition that comes from having delivered 40 projects and knowing what a schedule looks like two weeks before it slips.
What AI does is give the people who carry that judgment more room to use it.
That distinction — between work AI compresses and judgment AI amplifies — is also the most important lens you can apply to your next hire. The coordinator who executes a checklist is being replaced by technology faster than most firms have planned for. The person worth hiring in 2026 thinks clearly on incomplete information, makes good decisions under pressure and works alongside AI tools without needing someone to manage the process around them.
Firms still writing 2019 job descriptions are going to find themselves with a 2019 workforce in a market that has already moved.
The Order Matters—And So Does the Design
The AI conversation in commercial construction is happening almost entirely at the tool level. Which platform. Which integration. Which demo looked most impressive at the last industry conference. That conversation is not irrelevant. But it is starting in the middle.
Consider what the day actually looks like in most firms right now. A project executive is running three jobs simultaneously. The owner on Job A wants a schedule update. The answer requires calls to the super, the MEP sub, and the PM who is on-site at Job B. By the time the information is assembled, two hours are gone and the update is already 24 hours stale.
Meanwhile, a submittal on Job C has been sitting for four days waiting for an approval that requires the VP of operations—who didn’t know it was waiting because it was buried in an email chain he wasn’t copied on.
That is not a technology problem. That is a decision architecture problem. Who owns the schedule update? What is the trigger for escalation? Where does approval authority actually sit versus where the org chart says it sits? Until those questions have clear answers, AI produces better-formatted versions of the same confusion.
Now redesign that scenario with the operating model fixed first. Decision rights are clear. Escalation triggers are defined. The project executive knows exactly which decisions are hers and which ones require the VP. Information flows to the right person automatically because the system was designed to move it there.
In that environment, AI does something genuinely powerful—it compresses the time between information and decision. The owner update that took two hours now takes 20 minutes. The buried submittal surfaces before it becomes a delay. The schedule variance that used to show up in week three shows up in week one, when there is still time to act.
That is what AI-enabled decision architecture looks like in practice. Not faster chaos. Faster clarity. But clarity has to be designed first. AI accelerates whatever operating model you have. Build the right one, and the acceleration compounds. Leave the fragmented one in place, and AI gives you more data on a problem you still can’t resolve quickly enough to matter.
The firms that will look back at this period as a structural advantage asked a different question than their competitors. Not which tool to buy. But how work is actually designed inside the business—and what has to change so that AI makes decisions better and faster, not just more documented.
Three Questions Worth Asking This Week
You do not need another platform evaluation before you answer these.
Where are your best people spending time on work that AI could compress in the next 18 months—and what would they do with that capacity if it were freed? Where do decisions stall on your projects because the right information isn’t reaching the right person with the authority to act? And where is senior judgment—yours, your VPs’, your project executives’—being spent on problems that should have been resolved two levels below?
Those answers will tell you more about your AI readiness than any technology audit. They will also tell you exactly where the operating model work has to happen before the AI investment pays off.
McKinsey didn’t see the exposure in their own model until the cuts were already decided. You have the advantage of watching it happen first—and enough time to design around it rather than react to it.
That window is not staying open indefinitely.
Jane Gentry is founder of Jane Gentry & Company, a strategic operating model advisory firm working with mid-market CEOs and senior leadership teams in construction, manufacturing, logistics and professional services. She helps leadership teams clarify how work gets designed, how decisions get made and how AI creates leverage instead of accelerating drag. If these questions are live inside your business, a 60-minute diagnostic conversation is a good place to start. Reach her directly at jane@janegentry.com.
