IBM’s Watson supercomputer has beat Jeopardy champions, reconstituted recipes, and even helped create highlight reels for the World Cup. Now it’s taking on a new tech challenge; changing how the construction industry operates.

A new partnership between IBM and Fluor, a global engineering and construction company, will put the supercomputer’s computational skills to work on making building more efficient.

The new Watson-based system, in development since 2015 and now in use on select projects, will be able to analyze a job site “like a doctor diagnoses a patient,” according to Leslie Lindgren, Fluor’s vice president of Information Management. That degree of risk analysis, predictive logistics, and comprehension is no small challenge given the complexity of today’s construction megaprojects.

“These are multibillion dollar project sites, that are like walking into a city,” says Lindgren. “The sheer volume of data is tremendous.”

On especially complicated projects, Fluor will begin using two new tools, the EPC Project Health Diagnostics and the Market Dynamics/Spend Analytics, to make sense of the thousands of data points found on a crowded construction site. Constant analysis will help forecast issues before they show up, and automate how materials and workers are distributed. Eventually, the system will develop natural language functionality.

“Think of it as a living and breathing system,” says Sai Yadati, a partner for IBM Global Business Services. “This is truly one-of-a-kind for the marketplace.”

It’s a bold bet that reflects the ways in which the $10 trillion global construction industry continues to play catch-up with technology. That gap has created an opportunity for startups and technology firms to make a huge difference in an industry beset by increasing costs and labor shortages.

Applying artificial intelligence to the field won’t bring robots in hard hats, at least not anytime soon. But it can help with cost, scheduling, and quality for an industry in need of cost cutting. According to a McKinsey & Co. analysis, construction projects routinely lose up to a third of their value to waste.

“Construction is one of the least digitized industries, so many startups are seizing the opportunity to build technology that would increase efficiency within this market,” Michael Wholey, an intelligence analyst for CB Insights, told Curbed.

Startups sees opportunity in construction

The sharp rise in startup funding in the AEC (architecture, engineering, and construction) space makes it an ideal time to sell the promise of AI for the building trades. According to data from CB Insights, the industry saw $882.3 million in venture capital investment last year across 103 deals, and has already bested that in 2018, racking up $1.38 billion across 61 deals. KcKinsey found the construction tech industry has benefited from $18 billion in investment since 2013.

While AI’s more fervent converts predict massive leaps forward with adoption of the evolving technology, analysts predict construction will see more moderate gains.

Turns out the technology gap can’t be leapfrogged immediately; that same McKinsey analysis found that while AI could offer extensive benefits, “few E&C firms or owners currently have the capabilities—including the personnel, processes, and tools—to implement them.”

The report goes on to say that construction, in the short-term, will benefit from second-order improvements, especially as other industries in the supply chain continue to adapt the technology. Construction AI investment will still lag for the foreseeable future. Of the 12 industries McKinsey analyzed, 10 will increase AI spending at a faster pace than construction.

Carlos Serra, a managing director at JLL, a global real estate management firm, believes the impact of AI is, at best, five to 10 years off, due to the industry’s lack of standardization. JLL has made extensive bets in property technology, and recently built a new proprietary project management system that can work with AI applications. But he feels the fragmented nature of the industry makes it unlikely a technology will be introduced that works for everyone involved, from architects and owners to labor.

“We get compared to the auto industry, and people ask why we can’t be standardized like them, but it’s really not comparable,” Serra says. “For that to be fair, you’d have to have an individual design for each and every car. For us to get there, we’d need to make architecture redundant.”

Serra says that JLL is even using a basic AI system for a project in Los Angeles (he can’t name the client due to confidentiality issues). A system of small robots at the job site check and crosscheck real-word construction with a digital model of the project. While the technology can be helpful, it also added millions to the project, a premium Serra says won’t work for most job sites.

Betting on a more automated future

But that doesn’t mean there won’t be substantial investment. Fortune found many tech firms investing billions in construction tech firms, including Oracle, which purchased Aconex for $1.2 billion in February, and Trimble, which bought Viewpoint for $1.2 billion in April.

In addition, other ACE firms are creating their own systems. AECOM, the LA-based multinational engineering firm, introduced Capture, a new web and mobile platform that continuously analyzes construction sites and can build an accurate 3D model at any point in a project’s life-cycle.

Startup indus.ai, a construction intelligence-as-a-service solution which uses neural nets, computer vision, and analysis of millions of construction site photos to train AI to identify problems on existing job sites, just raised $3.7 million in seed funding.

With such a relatively small technology spend by the industry—McKinsey found engineering and construction only invested 1 percent overall into technology, significantly smaller than other industries—the opportunity for new technologies to change construction sites is vast.