There is no shortage of people willing to tell contractors what AI can do for them. Vendors pitch transformation. Conferences run panels on the future of construction. LinkedIn is full of posts about how AI will “change everything.”
Meanwhile, the average electrical or mechanical contractor is trying to staff three jobs at once, protect a margin that barely exists, and figure out whether any of this technology is actually worth the time it takes to evaluate it.
That disconnect is the real story. Not whether AI matters, but whether contractors can figure out which parts of it are useful before the next wave of pressure hits.
The labor problem is no longer a forecast
The construction workforce shortage has been discussed for years, but what’s happening now is different in scale and in consequence. The Bureau of Labor Statistics projects roughly 73,500 electrician job openings per year through 2032. Apprenticeship completion rates are not keeping pace. For every two or three experienced electricians leaving the workforce, only one or two replacements are completing training programs. Mechanical, HVAC, and plumbing trades are facing nearly identical math.
What makes this moment particularly painful is the collision between a shrinking labor pool and a historic construction boom. Data center construction alone represents tens of billions in planned U.S. spending over the next several years, and electrical and mechanical work makes up the majority of total cost on those projects. Add in semiconductor facilities under the CHIPS Act ($52.7 billion in federal incentives) and infrastructure work flowing from the $1.2 trillion IIJA, and the picture becomes clear. The problem for most contractors is no longer finding work. It is finding the people to execute the work they already have.
And when experienced estimators, project managers, and field leaders retire, they are not just leaving a headcount gap. They are taking decades of pricing knowledge, means-and-methods expertise, vendor intelligence, and project intuition with them. Most of that institutional knowledge has never been documented in any structured way. It lives in people’s heads, in scattered emails, and in spreadsheets on individual desktops.
Margins have no room left to absorb mistakes
Specialty contractors have always operated on thin margins. CFMA benchmarking data consistently shows net profit margins in the 2 to 5 percent range for most specialty trades, with general contractors often running even thinner. That has been the reality for years. What has changed is how many cost pressures are hitting at the same time.
Labor costs, typically 40 to 60 percent of project costs for MEP and specialty work, continue to climb. Material costs have plateaued well above pre-pandemic levels but remain unpredictable depending on the trade and the commodity. Insurance has become a particularly aggressive line item, with commercial general liability premiums rising significantly over the past several years driven by nuclear verdicts and broader market hardening. Workers’ compensation, fleet insurance, and the growing requirement for cyber insurance on certain project types have all added to overhead.
The result is a pattern that many contractors recognize even if they haven’t put words to it: revenue is up, but profitability is flat or declining. Firms are busier than they have been in years, but the work is not translating into stronger financial performance. When your margin is 3 percent, there is very little room for a bad estimate, a staffing miscalculation, or a material cost surprise.
What AI adoption actually looks like in contracting right now
If you want an honest picture of how AI is being used in contracting today, start here: the most common use is individual employees quietly using ChatGPT and similar tools on their own, without any formal company policy, training, or awareness from leadership.
Project managers are using it to draft RFI responses. Estimators are using it to summarize specifications. Business development staff are using it to write proposal narratives and emails. Safety managers are generating job hazard analyses. This informal, unmanaged adoption is happening across companies of every size, and most owners and executives have no visibility into it.
On the formal side, the AI category with the clearest traction in contracting is estimating and quantity takeoff automation. Tools that use computer vision to read blueprints, identify spaces, count fixtures, and extract quantities are helping firms bid significantly more work with the same estimating staff. That matters because experienced estimators are among the hardest roles to fill and the longest to develop.
Beyond estimating, there are promising applications in safety monitoring, progress documentation, and knowledge management. But the honest reality is that most of these tools are being used by a small percentage of firms, primarily the largest and most technology-forward contractors. The vast majority of specialty contractors, particularly those under $50 million in annual revenue, have not taken any meaningful action on AI.
The real barriers are not about technology
When contractors hesitate on AI, it is rarely because they don’t believe the technology works. The barriers are more practical than that.
Construction data is messy. It lives in different systems that don’t talk to each other. Estimating might run on one platform, project management on another, accounting on a third, and critical operational knowledge in spreadsheets and email threads. AI tools need clean, structured data to deliver value. Most contractors simply don’t have it.
Workforce culture matters too. Superintendents, foremen, and senior field staff are often the most resistant to new tools, and they are also the people whose buy-in matters most. When a company has already cycled through two or three technology rollouts that promised transformation and delivered friction, asking the team to try something new is a harder sell than any vendor demo suggests.
And there is the practical question of capacity. Most specialty contractors under $50 million in revenue have no dedicated IT staff. There is no one internally to evaluate tools, manage implementation, or troubleshoot problems. Technology decisions end up being made based on trade show conversations and peer recommendations, not systematic evaluation.
The competitive landscape is shifting underneath all of this
While contractors work through the AI question, the broader competitive environment is changing in ways that will reward firms with better positioning, stronger visibility, and more deliberate business development.
Design-build delivery continues to grow as a share of nonresidential construction, and DBIA research shows it approaching half of all nonresidential spending. That shift changes what winning looks like for specialty subcontractors. Design-build rewards earlier engagement, preconstruction capability, and relationships built before the bid hits the street. Contractors who can only compete in traditional hard-bid environments are losing ground.
Private equity has accelerated the pressure. Platforms like Comfort Systems USA, APi Group, and IES Holdings continue aggressive acquisition strategies in specialty contracting. That means well-funded competitors entering local markets with deeper resources for talent, technology, and bonding capacity. At the same time, the majority of construction company owners are approaching retirement age, and a relatively small percentage have formal succession plans in place. The combination creates both risk and opportunity depending on where a firm stands.
What the contractors getting this right have in common
The firms that are positioning themselves well right now are not chasing every new tool. They are doing a few things deliberately.
They are using AI where it has already proven useful, particularly in estimating and preconstruction, to get more out of the staff they have. They are putting basic policies in place around how employees use AI tools, so that the informal adoption already happening is at least guided rather than invisible. They are starting to document institutional knowledge before their most experienced people leave. And they are thinking more seriously about how they show up in the market, because in a world where every contractor is busy, the ones who can clearly communicate their value and track record are the ones who will keep winning the right work.
None of that requires a massive technology budget or a dedicated innovation team. It requires clarity about what actually matters and a willingness to act on it before the window closes.
The real divide forming in contracting right now is not between companies that use AI and those that don’t. It is between companies that are building the habits, the data discipline, and the operational foundation to adapt, and those that are too busy working in the business to work on it.
That gap is going to get harder to close, not easier.