By Chuck Gallagher | Business Ethics Keynote Speaker | AI Speaker and Author
TL;DR: AI construction workflows are quietly reshaping how contractors win bids, manage projects, and respond to owners — and firms ignoring this shift are losing ground without realizing it. Chuck Gallagher, AI speaker and author, argues that the real ethical question is not whether to adopt AI, but whether leaders are honest about what gets compressed out of the workflow. Speed without judgment is the same trap that has produced every white-collar scandal of the past forty years.
Five hours a week. That is the number a recent Birmingham Group piece asks construction executives to consider. If every project manager, estimator, and operations leader in your firm could reclaim five hours a week by handing documentation work to AI, what would change about your competitive position over the next twelve months? It is a fair question. It is also the wrong place to stop the conversation.
The construction industry sits at an interesting intersection. It operates under tight margins and tighter deadlines, and it has historically been slow to adopt new technology. As a business ethics keynote speaker who has worked with leadership teams across heavily regulated industries, I have watched what happens when an industry that has resisted change suddenly embraces a tool that promises to make everyone faster. The promise is real. The risk nobody is talking about is also real.
The Birmingham Group’s reporting is accurate. Preconstruction teams are using AI to generate scope summaries from large document sets. Estimators are running bid package reviews faster. Project managers are producing RFI and submittal summaries in a fraction of the time. McKinsey & Company research cited in the article shows generative AI can improve productivity in knowledge-heavy roles by 20 to 40 percent. The Bureau of Labor Statistics JOLTS report continues to show elevated job openings across construction management, and Associated Builders and Contractors has projected workforce shortages through the rest of the decade. Nobody is hiring their way out of this. The firms that pull ahead will be the ones that compress the right hours.
Here is where I want to push back on the framing. The article describes AI adoption as the elimination of friction between receiving information and acting on it. That is a useful description for a software demo. It is a dangerous one for a construction executive. A great deal of what looks like friction in a project manager’s week is actually judgment. When an experienced PM reads a 200-page submittal, the time spent is not just processing words. It is pattern recognition. It is the seasoned eye catching a subcontractor’s hedge in section 14 that does not line up with what was promised in section 3. AI summaries will catch most things. That is precisely the problem.
What gets lost when documentation gets compressed?
There is a phrase from my CPA days I think about often: management override of controls. It refers to the moment when a senior person, under pressure, decides the standard process is too slow and works around it. That moment is the seed of nearly every accounting fraud I studied during my federal sentence and every one I have studied since. The technology changes. The pattern does not. When AI compresses a documentation cycle from eight hours to forty minutes, the question nobody asks is what was being learned during those other seven hours. Sometimes the answer is nothing useful. Sometimes the answer is the very thing that prevents a billion-dollar mistake.
I am not arguing against AI adoption in construction. I am arguing for a specific kind of leadership discipline as it happens. The Birmingham Group is right that integration discipline is the differentiator. What I would add is that integration discipline has an ethical dimension the article does not name. Every workflow you compress is a workflow where mistakes will surface later, often in front of an owner, an insurer, a regulator, or a jury. Construction already deals with structural defect litigation that runs into the billions of dollars annually. Faster documentation that papers over weaker review will produce more of that, not less.
Who is accountable when an AI summary misses what a human would have caught?
This is the question construction executives need to answer before they expand AI use, not after. If a project manager signs off on an AI-generated RFI response and that response contains a subtle error costing the owner three months of schedule, who owns that mistake? The PM? The vendor that built the AI? The executive who pushed adoption? In every ethics failure I have studied, accountability blurred before the failure became visible. Speed first, governance later, consequences last. That is the order of operations that produces scandals.
As an AI speaker and author who works with leaders on governance frameworks, the recommendation I give every client is the same one I would offer any construction firm reading the Birmingham Group’s piece. Before you expand AI use across your workflow, write down three things. First, what is each AI tool authorized to do, and what is it explicitly not authorized to do. Second, who in your firm is personally accountable when an AI-assisted output causes a problem on a project. Third, can the people affected by your firm’s AI-assisted decisions understand how those decisions were reached and challenge them if they are wrong. If you cannot answer those three questions in writing, you are not ready to scale. You are ready to get burned.
The construction firms that win the next decade will not win because they adopted AI faster than competitors. They will win because they kept the human judgment that protects their reputation and their clients while letting AI handle the genuinely repetitive work. The 2,500 hours per team that the Birmingham Group cites is real. What you do with those hours is the actual leadership question. Spend them on relationships, problem-solving, and ethical review. Do not spend them on more pursuits at lower quality. Every choice has a consequence, and the consequences of cutting corners in construction tend to outlive the executives who cut them.
Frequently Asked Questions
How are construction companies actually using AI in 2026?
Construction firms are using AI primarily for documentation-heavy tasks: generating scope summaries during preconstruction, reviewing bid packages, drafting RFI and submittal responses, preparing safety documentation, and accelerating proposal development. According to research from McKinsey & Company cited by The Birmingham Group, generative AI can improve productivity in knowledge-heavy roles by 20 to 40 percent. Field work like concrete pouring or steel erection is largely untouched by AI in 2026.
What is the biggest ethical risk for construction firms adopting AI?
The biggest risk is the gradual erosion of judgment that happens when documentation cycles get compressed. As I have argued at ChuckGallagher.com, every ethics failure follows a predictable pattern of need, opportunity, and rationalization. When AI summaries replace the slow read of a submittal, experienced reviewers stop catching the small inconsistencies that prevent large disputes. The risk is not dramatic disruption. It is the silent loss of the careful eye that protects owners, subs, and the firm itself.
Will AI replace construction project managers or estimators?
No. AI supports construction professionals but does not replace the judgment that experienced people bring to contract interpretation, owner communication, field leadership, and risk evaluation. The Bureau of Labor Statistics JOLTS report continues to show elevated job openings across construction management roles, and Associated Builders and Contractors has projected ongoing workforce shortages through the remainder of the decade. The industry needs more capable leaders, not fewer.
What questions should construction executives answer before scaling AI use?
Before expanding AI across the workflow, executives should answer three questions in writing. What is each AI tool authorized to do, and what is it explicitly not authorized to do? Who in the firm is personally accountable when an AI-assisted output causes a problem on a project? Can owners, subcontractors, and employees affected by AI-assisted decisions understand how those decisions were reached and challenge them if they are wrong? Without clear written answers, the firm is exposed to liability and reputational risk that no productivity gain offsets.
How does the construction industry’s AI adoption compare to other regulated industries?
Construction sits at a midpoint. It is moving faster than utilities and slower than financial services. The challenge for construction is that errors tend to manifest later, sometimes years after substantial completion, and they often surface in litigation rather than in real-time monitoring. That delay between mistake and consequence makes governance discipline especially important during AI adoption, because the warning signs of a problem may not appear until well after the responsible parties have moved on to other projects.
I would like to hear from construction executives, project managers, and owners on this one. Where in your workflow has AI already saved meaningful time, and where do you think it has quietly degraded the quality of your reviews? Share your experience in the comments below — the conversation in this industry is only just starting, and the firms that learn fastest from each other will be the firms that get this right. To help you think through your own situation, here are five questions worth sitting with.
Five Questions for Further Thought and Consideration
- When AI compresses a documentation cycle from hours to minutes, what specific patterns of judgment are at risk of being lost in your firm?
- If an AI-assisted RFI response contained a subtle error that cost an owner three months of schedule, can you name today the person in your organization who would be personally accountable?
- What standards does your firm apply to the AI tools embedded in your document control, scheduling, and bid management software, and who reviews those standards annually?
- How are you measuring the quality of AI-assisted work, not just the speed, and how does that quality measurement feed back into your hiring, training, and promotion decisions?
- If a subcontractor or owner asked your firm to explain how a particular AI-assisted decision was reached on their project, could you give them a clear answer they would find satisfying?
Related Articles:
Cyber Security Report 2026: AI Is Now an Ethics Problem
AI in Education 2026: Why Ethics, Not Apps, Will Decide Who Wins
