AI Adoption in Construction: Why Teams Resist and the ADKAR Framework That Works

After years of building scheduling and risk management software, I have watched every wave of new technology hit the construction industry. A year ago, I tried integrating Cursor with Claude Code on my machine, and it was like having a really bad intern. I re-upped it a month ago, and it was amazing.

In my lifetime, there has not been this big an investment this quickly in any given technology, and I believe AI holds the potential to be more transformational than the internet. The problem is that it is landing in an industry where a new idea is not a good idea. Closing that gap requires real change management in construction.

TL;DR

Construction teams resist AI because of professional inertia, fear, and conservatism baked into contracts and culture. The ADKAR model, a proven approach to change management in construction, gives you a structured path through that resistance. Start by showing your team what AI removes from their plate, not what it adds.

The Current State of AI Adoption in Construction

A year ago, I tried integrating AI coding tools into my workflow, and it was like having a really bad intern. I went back a month ago, and it was so much better. The models are evolving so quickly that it's genuinely difficult to judge where the technology stands at any given moment.

The most immediate opportunity for AI in construction project management is natural language access to project data. A CFO or contractor could ask "how many activities have less than two days of float right now?" and get a direct answer without needing to be a scheduling expert.

Beyond that, always-on agents could flag near-critical activities before they turn red, rather than waiting for a human to catch them in a bi-weekly review.

The bigger barrier isn't the technology, it's the industry. Construction treats new ideas with suspicion, and that inertia is baked into contracts and reinforced by a forensic scheduling expert witness industry built around the status quo. A lot of the industry isn't even aware they need to change yet.

Awareness will come from economic pressure. Someone will figure this out, start undercutting competitors, and the rest will follow. The fastest path to internal buy-in in the meantime is the WIIFM, or asking ‘What is it for me?’.

Slide on the current state of AI adoption in construction. — AI is advancing quickly, but construction adoption is shaped by industry inertia, contractual habits, and growing economic pressure to work more efficiently.

Why Construction Teams Resist New Technology

The resistance is not irrational. It comes from real, structural places. Here are the six sources I see most often.

Source of Resistance	What It Looks Like
Contractual inertia	Contractually, many of the contracts require a P6 schedule.
Professional inertia	You have to be an expert scheduler and an expert at P6 to actually navigate it, which then makes that person special.
The litigation industry	The whole forensic scheduling expert witness industry, a multi-billion dollar industry, they are invested in knowing how to fight over a P6 schedule.
Fear of job loss	One thing that AI generally as an industry is not talking too much about is the fear that people have that AI is going to take their job.
Safety conservatism	There are some good reasons for conservatism in engineering. You do not want the bridge to be under-engineered.
Scar tissue	22 years ago we built a daily reporting Palm app and they would not do it. It fell under the truck wheels. They wanted their handwritten book.

In my experience over many years, a new idea is not a good idea in construction and engineering. Understanding these sources of resistance is the first step to getting past them.

What Is the ADKAR Framework?

ADKAR is a change management model with five sequential stages: Awareness, Desire, Knowledge, Ability, and Reinforcement. It was developed by Prosci founder Jeff Hiatt from research across 1,000+ organizations, focusing on individual change because organizational change only happens when individuals change. It is one of the most effective approaches to change management in construction because it diagnoses exactly where adoption is stuck.

I am a big proponent of organizational change management as an approach to managing change.

ADKAR stands for: Are you aware that you need to change? Do you have a desire to change? Do you have the knowledge, the ability, and are you being reinforced? You have to get through each stage gate to actually have a successful change.

Applying ADKAR to AI Adoption in Construction

Each ADKAR stage has a gate. If your team is stuck, look for the gate they have not passed through.

Step 1: Awareness

Awareness means your team understands that the current way of working has a measurable cost. A lot of the industry is not even aware that they need to change. Until they are aware, nothing else matters.

Show them the numbers: what duplicate schedules cost, how many hours go to manual updates. Awareness will come from economic pressure. Somebody is going to figure this out and start undercutting other people. You can wait for that pressure or create it internally.

Step 2: Desire

Desire means your team personally wants the change. Find what I call a WIIFM, which is what is in it for me. If a PM's worst nightmare is updating spreadsheets, help them automate that. They will like AI because we have done something for them, versus training AI to take their job. It would emancipate an experienced PM from low-level work and allow them to apply their time to higher-value functions.

Step 3: Knowledge

Knowledge means your team knows how to use AI in their specific workflow. Interrogate the data set, the schedule itself, using AI. Use it to learn more about your own schedule, to instruct yourself. But always verify. Do not just let the agent run wild.

Step 4: Ability

Ability means the tools are accessible enough for non-specialists. Natural language querying is going to level the playing field for all the professionals who have questions, whether it is the CFO, the project sponsor, or the contractor. You do not need to know P6.

Step 5: Reinforcement

Reinforcement means the change sticks. If big contracting organizations were to adopt a platform and a standard, you would see it drive a lot faster. Build it into your project cadence: make AI-generated reports the default, and the old way becomes harder than the new way.

ADKAR framework showing five stages of change management. — The ADKAR framework helps construction teams identify where AI adoption is stuck and what needs to happen before change can take hold.

Where AI Delivers the Fastest Wins

The best way to build momentum through the ADKAR stages is to target AI at the work people already hate doing. These use cases generate the fastest adoption because they remove pain and do not add complexity.

Use Case	Why It Works
Automated reporting	So much project management is driven by filling out forms, updating spreadsheets, all that repetitive mindless work. Automating this frees PMs for higher value work.
Natural language schedule queries	Build a data lake off a whole project and then natural language query that project for answers. No P6 expertise needed.
Portfolio-level visibility	If you are a project controls leader with 30 or 40 projects around the world, have the whole vision available to you to query against from finance through scheduling through safety.
Agentic monitoring	Build a specific agent to watch for schedule progress. Is something near critical but trending towards critical. Agents catch it before it turns red.
Continuous risk analysis	Have AI query to see if any of those risks are emerging and update them and rerun the Monte Carlo. You can suffer from death by a thousand cuts.
Dashboard automation	The amount of time I was able to save just having the system manipulate the data set and then update the dashboard was terrific.

If your AI use case adds work for the end user, adoption will fail no matter how good the technology is.

The Future: What Construction Looks Like in 2035

The pace and scale of AI funding are unlike anything construction has seen before. As transformative as the internet has been to the world, I think AI holds the potential to be more transformational, and it is better funded. The question is not whether AI will transform construction. It is how fast.

Communication over data chasing

According to the PMBOK, 90% of a project manager's job is communication. I agree with that. If we could use AI to better inform a project manager as to project status, trends, and change orders, then they can do a better job communicating and executing. Because if the job gets out of sync through lack of communication or lack of data update, then you are driving blind.

Automated progress tracking

Progress is currently a human activity. I do not think it needs to be in 2035. We should be able to progress the project based on sensors, or by 2035 maybe there will be a hundred robots on the site reporting the data back in real time. You could run a digital twin in advance. Build the project and say, on the fifth floor the robots are falling over, so let us fix that before we start actually building anything.

Schedulers as strategists

The real professional schedulers that I have met are craftsmen. They understand what they are trying to accomplish and how to look at the model. They are not just clicking keys. That is where you can get more value out of their experience and their intuition.

Start Driving AI Adoption on Your Projects

Change management in construction comes down to diagnosing where your team is stuck in ADKAR, solving for that specific stage, and building momentum by targeting the work people already hate. They will like AI because we have done something for them versus training AI to take their job from them. Start there.

Explore Mastt's AI in construction research or see how AI project management works in practice.

Topic:

AI in Construction

Written by

Timothy Mather

Tim Mather is a former CTO and COO of PMA Technologies, where he spent 23 years developing construction scheduling and risk analysis software. He co-developed NetPoint® and NetRisk™ and co-authored "Core Traits of a Reliable Schedule." At Mastt, he writes on construction scheduling, project controls, and AI in construction.