Author: David Wilson, Chief Innovation Officer, Bechtel 

The idea that Artificial Intelligence (AI) and robots will take our jobs and eventually conquer the world makes for a great summer blockbuster, but this fear of AI-robot overlords is grossly exaggerated. 

I personally welcome the AI and robot revolution. Robots won’t replace our talented construction craft. Instead, a new generation of robots will strengthen our builders by performing highly repetitive, monotonous, hazardous, and less-productive tasks. In the same sense, AI – which performs decision-making tasks traditionally reserved for humans – won’t render our knowledge workers irrelevant. AI will allow us to better predict outcomes, design complex projects, and automate day-to-day decision-making tasks.

The AEC industry operates in a sea of risk. The usual suspects, like incremental weather, material and labor shortages, design challenges, or regional strife are not uncommon. Extreme and hard to predict events often occur. No matter how detailed the plan or thorough the risk register, something unforeseen and highly impactful will happen.

Due to these uncertainties, it is rare to find projects similar enough to establish meaningful, structured correlations. There is one Hoover Dam, one Ivanpah Solar Field, one Corpus Christi LNG Program, one Jubail Industrial City, one Chunnel – the list goes on. Predicting future outcomes due to unpredictable events is complex and laborious. This is where AI can assist project delivery, by helping us gain insight from hidden correlations. Exposing the relationships found in disparate and unstructured data sets will produce more efficient and optimized schedules.

Using current practices and technology, only a handful of schedule permutations can be fully evaluated. Manually scripting sequences is time consuming and often limited by the knowledge of the individual planner. However, in the future, a planning process employing AI will allow our planners to quickly create and evaluate millions of scenarios, even for schedules with thousands of activities. Possessing this array of scenarios, a project team can select the sequence to best optimize project outcomes for owners and operators.

By first reviewing Bechtel’s disparate project data from the last 10 years, an AI-powered assistant will gain a refined understanding of schedule correlations, as well as improve its prediction capabilities. And 10 years is only the beginning, eventually we will be able to train AI to interpret our 120 years of historic data. More than a few start-ups, for example Alice Technologies, and large enterprises, like Bechtel, are actively working to solve these types of project planning problems.

It’s time we put AI and robotic solutions to work and free our designers, buyers, and builders to tackle new challenges. With trillions of dollars of infrastructure spending needed to keep pace with projected growth, we need to leverage every available advantage and technology to productively build the world. AI can be adopted in AEC, but off the shelf applications have to be individually optimized for each project.  At Bechtel, we’re continuously working to accelerate the transformation of the built world and resilient infrastructure through applied experience, technology, and innovation to enhance value for our customers, teams, and the general public.