Q&A: What it Takes to Make AI Stick in Construction

Pandya: Because each project in our industry is essentially a standalone effort with unique teams and limited connectivity, the perceived risks of experimenting with new tech can often outweigh the perceived upside. There’s a reason we’re risk-averse and a low-margin industry. People worry that, if they get it wrong, they might lose their entire margin on a job, whereas if they get it right, the improvement may seem incremental. But that same filter is a feature: any solution that breaks through has genuinely earned it and tends to be sticky.

Hegde: Construction likes simplicity. Anything that creates friction by adding work or forcing a change to workflows faces resistance to adoption. Construction wants solutions where you add value without making people change their habits, because otherwise deployment at scale becomes difficult. We call this concept gentle disruption. In practice, this means the most successful AI tools tend to be those that complement existing processes rather than reinvent them.

Hegde: AI literacy is a problem, but not such a massive problem because people are using AI in every part of their lives. Tools like Google Maps and social media already rely on AI behind the scenes. That said, there is still a lag in digital literacy in construction. It’s changed dramatically since COVID, but it's still uneven. The second issue is the fear that AI or robotics might displace jobs.  If that’s the mindset we have, then [lack of] literacy is not the problem—fear is the problem. It’s a valid concern and the way jobs look in the future will be different than they do today.  We have to find a balance there and do it the right way. Leadership must reassure teams that AI is meant to augment their work, not replace them, while also helping them build new skills for the future.

Pandya: I don’t believe that finding the right tool alone is an answer for any firm. It’s more important to develop the organizational muscle of identifying a problem the team is motivated to solve, scanning the ecosystem efficiently, finding three to five solutions to pilot, and structuring a process for scaling. Much harder, but that organizational muscle is actually the biggest competitive advantage you can ever have. That repeatability compounds. A silver bullet is a one-time event; organizational muscle is a durable edge.

Pandya: I don't think anyone knows the true unit economics of AI yet, and current pricing is masking that. Many offerings are priced for adoption, not margins, and the foundation models underneath them are still finding their cost structure. The real test comes when habits have formed, pricing resets, and model selection tradeoffs between cutting-edge and open-source start showing up in the margin. That's when every use case faces the vitamin or painkiller test: Does the value hold when the discount disappears?

Hegde: AI is going to be like electricity. It’s going to be everywhere and in everything we do. AI will eventually become a normal part of construction, embedded invisibly in many processes. But the industry’s core mission will remain the same. We’re still trying to solve our biggest problems; AI is just a tool to help us do that much more efficiently than we could in the past. The effort and judgment professionals bring to projects isn’t replaced by AI—it’s augmented. We do need to increase our understanding of how to adopt AI responsibly (what the risks are, what the benefits are) and educate our people accordingly. Success with AI will come from keeping it purpose-driven, demystified and human-centered.