Construction is entering one of the most consequential periods in its modern history.
The global AI boom is triggering one of the largest infrastructure buildouts in decades. Global political pressures are pinching resources and changing regulatory requirements. And in the midst of it all, the US is facing a projected labor shortage of 349,000 workers in 2026 alone.
For years, many construction companies have been able to absorb inefficiencies through schedule extensions or increased costs passed downstream. But today, every delay, every rework, every wasted resource, matters.
It’s why construction companies are being forced to rethink how work actually gets done in the field and how technology can support faster and more predictable delivery. Yesterday’s tools and ways of working can’t deliver speed and scale today.
But there is one barrier to efficiency that companies have entertained for far too long: Pilots.
The Pilot Problem: Built for the Back Office
The Wall Street Journal recently reported that companies are actively moving away from the word “pilot” when talking about AI initiatives. Mentions of the term on earnings calls dropped significantly in late 2025.
It makes sense. An MIT study found that 95% of enterprise AI pilots fail to create measurable financial impact. The word itself has become associated with stalled progress and a lack of commitment — two things the construction industry can’t afford right now.
For most, pilots are well-intentioned but often an excuse for companies to be half in, half out. Often, the organization is running multiple pilots across the organization but rarely emerging with insights to move to real implementation. Instead, they’re testing repeatedly. I call this a case of “pilot-itis.”
Pilots tend to create more operational drag than efficiency, especially around AI and infrastructure. Success depends on whether field workers adopt the tools consistently in live environments, under real pressure to complete jobs safely and on time.
On top of that, pilots are designed around the needs of the back office, not the realities of the field. Consider the procurement process. If a pilot is successful, the tools are actually taken away from field workers for several weeks to go through procurement. In that time, field workers are expected to revert to pen-and-paper methods.
After the tools clear procurement (assuming they do), we ask field workers to switch back to the tools, disrupting their workflows not just once, but three times. Over time, this creates fatigue, skepticism and disengagement. We’ve seen field workers simply wait for the pilot to end so operations can return to what they’ve always known.
In this time when companies are looking to lower risk and increase efficiency, pilots are not the answer. How can we move beyond them?
Beyond Pilots: Lessons from Infrastructure Deployments
Instead of pilots, we encourage our customers to deploy and scale two or three big projects and grow from there.
For example, when my company partnered with infrastructure services provider Amey across its Highways division, our goal was not simply to “test AI” and see the results. We wanted to understand how frontline intelligence, meaning bringing real-time insights to field teams while they work, could operate inside real field conditions to solve operational challenges teams were already facing every day.
What does that look like in action? We’ve learned a few lessons about what it takes to move beyond pilots and drive adoption:
- Know when testing is warranted. For example, if other companies in your industry are deploying solutions that make them more successful, follow their lead. But don’t just launch another pilot. (Let’s be honest: How often do you really need to prove something?) Follow quickly, buy and scale and reflect on your learnings.
- Set a clear start and stop point. Too many pilots become open-ended experiments with no urgency behind them. Define the timeline and specific outcomes you’re measuring and make decisions quickly. Test intentionally. Instead of running endless pilots and losing time, resources, and steam, choose a few things you believe will change the game.
- Validate with a meaningful group, not the entire organization. The objective shouldn’t be to run a small pilot for a long time; at the same time, successful transformation doesn’t start with 10,000 workers at once. You might start with a few business units that you know are welcoming of new ideas and ways of working and grow from there once everyone is using and learning from the tech. By the end of our deployment with Amey, 85% of users were consistently active on the platform.
- Solve real operational problems, not artificial test cases. In highways operations, we focused on addressing challenges like avoidable job aborts, incomplete site information, inconsistent preparation and resource waste. Technology adoption happens when field teams see direct value in their day-to-day work.
- Collect enough data for actionable insights. AI systems are only as valuable as the operational context behind them. Running limited tests with minimal field engagement might produce activity, but it doesn’t make back-end and field teams more intelligent. The real leap comes from working with partners that draw on cross-industry data and hard-won experience. That’s where you move from incremental improvement to exponential insight.
- Protect momentum after deployment. One of the fastest ways to kill adoption is forcing field workers backwards into legacy processes midway through implementation, like in the procurement example earlier. If teams have already integrated new tools into their workflows, companies should avoid unnecessary disruptions that require workers to stop, restart, and relearn processes multiple times. Make the change once and support workers through it properly.
Construction companies need operational commitment to the tools and processes that make field work better, not more experiments.
Technology absolutely has the potential to reshape construction for the better, but only if companies rethink how they introduce change in the first place. The construction industry must move beyond pilots and focus instead on supporting field workers with tools that improve how work gets done. Because when the frontlines experience improvement, the back office sees the results.
