If You Can Win The Supers Over, You Win The Company

That line came from a conversation I had with an implementation manager at a VC-backed construction AI startup. He has spent years rolling out software to large organizations in construction and his take really stuck with me. The hardest part isn't the technology but the behavior change.

Over the past year I have had many conversations with folks who are actively building AI tools for construction. Implementation managers. Founders. Engineers. Product managers.

What stands out is that these people aren't talking about robots replacing carpenters. They're talking about documents. Communication. Change orders. The daily friction of tracking decisions and finding information. These are the unsexy problems that eat up hours every week but don't always make for flashy conference demos.

The common thread I observed across all of these conversations, is that AI adoption needs to be focused on the people in the field using these tools rather than coming from the top down.

The Implementation Gap

The conversation around AI in construction seems to mostly focus on which tools exist and what features they have. But the real bottleneck seems to be implementation.

A recent RICS survey found that 45% of construction organizations have zero AI implementation at all. Only 12% use AI regularly in specific workflows. In other words, the "AI revolution" hasn't reached most jobsites yet.

For most field teams, AI represents a bigger shift than going from spreadsheets to Procore. We are asking people to rethink how we engage with software and technology. In my own experience as a construction PM, I often found digital tools to be more of a nuisance than an aid. Forcing myself to document conversations, write reports, and flag issues before they arise. I think the BIGGEST shift with the coming wave of technology, is going to be more about implementation than about the tools themselves.

What I'm seeing is that successful AI implementation in construction comes down to a few things:

  • Showing value without adding complexity. Field teams already feel daunted by the thought of adding another piece of software. Another platform that takes six hours to learn isn't going to get used. The tools that work are ones that feel like a layer on top of existing workflows, not another software claiming to automate everything.

  • Speaking the language. The people rolling out these tools need credibility with field crews. That means understanding jobsite realities, not just technical capabilities.

  • Making the tool proactive. The most promising approaches I've seen involve AI that runs automatically and delivers outputs without the user needing to do anything except read the results. That makes the value obvious.

This is where I think the opportunity lies for smaller contractors. Big enterprises move slow on implementation. A 15-person design-build firm can actually adopt faster if they understand their options and find tools that match how they already work.

Why The Next Wave Is About Adoption

In fact, I think by definition every wave is about adoption but it's a different way of looking at it. The rate of change and improvement with AI technology is dramatically higher than the rate of adoption. Put another way, if we ceased to see any improvement of AI models and were only left with the ones currently available, humans and businesses still have a lot of catching up to do in order to maximize the benefits of current tools.

Rather than feeling overwhelmed by the constant arrival of shiny new tools, I think businesses are going to see more value if they pause and start focusing on adoption instead. The companies getting traction are focused on document intelligence and voice-first field tools. Chatting with your specs instead of scrolling through PDFs. Dictating site notes instead of typing them into an app at the end of the day. Asking a question about a submittal and getting an answer in seconds instead of hunting through emails.

This is the practical AI that's actually reaching jobsites in 2025. It's not flashy, but it solves problems contractors deal with every single day.

I've been testing some of these tools myself. Voice capture for site documentation. AI that structures messy notes into professional reports. Chat interfaces trained on energy codes and product manuals. The technology works, so the question remains: will people actually use it?

That brings us back to implementation. The best tool in the world is worthless if it sits unused. And getting construction professionals to change how they work requires more than a good demo. It requires trust, credibility, and a genuine understanding of what the job actually looks like.

One Thing You Can Try

Next time you're scrolling through a PDF on your phone looking for an install spec or code requirement, notice how long it takes. That's the friction these tools are trying to eliminate. If you want to experiment, try uploading that same PDF to Claude or ChatGPT and asking it your question instead. It's not perfect, but it's a simple way to feel the difference firsthand.

I'm implementing these tools with a few contractors right now. If you want to try one on a real project, send me an email.

What's your biggest question about AI implementation right now? Drop me a message!

Murray

References:

Keep Reading

No posts found