By KERRY SMITH BUCK

The future of construction jobsite productivity is driven by automation and augmentation, not by replacement.

These words came from a featured break-out session at AECTechCon™ in St. Louis last week. Session presenter Nicholas Tilford –  Procore’s industry transformation manager for North America – told the audience that what bothers them about their day-to-day jobs now probably won’t be an issue in five years’ time, thanks to the advancement of AI.

“Either you’re going to use AI to attack the things you love, or you’re going to leverage it to accomplish the things you hate doing,” Tilford said. “It’s a sure thing: technology will augment workers, not replace them.”

Tilford’s session was one of two presented by Procore at AECTechCon 2026, a premier two-day conference hosted by the AGC of Missouri, which took place May 6–7 at the St. Charles Convention Center in St. Charles, Mo. The conference brought together more than 400 architecture-engineering-construction professionals to focus on digital transformation, technology integration, BIM/VDC and industry innovation.

As it stands today in the construction industry, 18 percent of project time is spent searching for data, Tilford said, referencing Dodge Construction Network’s Future State of Construction five-year outlook.

“And 28 percent of a project’s total time is spent on rework or rectifying issues,” he added. “A full 42 percent of leaders (who responded to the 2026 Dodge survey) say upskilling has one of the biggest impacts on digital transformation. Forty-five percent of respondents said they want to develop technical skills.”

Starting with the finish line and working back from there is a smart strategy for gaining the best, most comprehensive – and usable – intelligence from open-source platforms such as Claude and Gemini, Tilford said. “You provide the goal or output and the AI agent takes over and applies reasoning,” he said. “For example, connecting your apps and sites – such as your calendar and Gmail – to the larger context of your asks allows AI to pull from that tool.”

Starting at the end and working backward allows individuals – such as project managers and superintendents – to better prepare for what the answer is, according to Tilford. “Setting your end goal can have a multiplier effect,” he added.

Remembering that the human element is just as important as the systems that construction professionals apply is key, Tilford says.

“A large language model (LLM) provides the multiple AI factor, whereas an AI workflow does not,” he noted.

An LLM is often considered better than a static AI workflow when flexibility, reasoning and context-aware generation are required – rather than rigid, repeatable steps. While workflows follow fixed rules, LLMs adapt to unpredictable inputs, making them superior for creative tasks, complex understanding and conversational interfaces, Tilford says.

“Imagine what you can do with all the data that you already have about your projects,” he said. “Now imagine that you’re compounding that data, month after month. That one photo that you took on the jobsite…86 percent of the time, it’s never used again. But when you’re leveraging AI, the quality of the data matters. Everything we do going forward is a history-teaching event.”