How AI is Reshaping the Future of Construction

Tony Peters, VP of Architecture at Gray AES, remarks on the change. “My standpoint is that AI isn’t going to go away, so I’m going to be an early adopter and a change agent, and I’m going to teach myself everything I can to stay relevant.”

This article explores AI’s role in shaping the future in the construction industry: how its uses can affect construction management as well as what makes AI tools and technology distinct from earlier waves of digital innovation.

Machine Learning Improves Risk Mitigation and Construction Site Awareness

The ability to identify patterns from large datasets is a core function of machine learning. With enough historical and real-time data, algorithms can detect early indicators of risk or inefficiency.

Says Peters, “Using automation for jobsite walks is a great way to allow people around the world to visit the site virtually without having to visit physically.” An Engineering News-Record (ENR) article showed camera and AI capabilities from EarthCam that use panoramic images of a job site and analytics to identify on-site conditions such as safety status and material availability, and assess readiness based on factors such as weather and schedules.

AI pattern recognition tools are like having highly experienced managers overseeing multiple aspects of a project, looking out for signs of trouble and pitfalls to avoid. Scheduling and safety oversight solutions are an invaluable way to help teams focus on potential trouble spots before they become serious problems.

Peters is looking at the benefits of using AI for risk mitigation. “We can load a database with what the potential risk indicators are and the mitigating tactics on how to solve the problems. We could then use AI to help project teams try different solutions before we have to elevate issues to executive levels.”

Generative AI Accelerates Project Planning and Exploring Options

Striving to replicate the work of experienced people, generative AI creates new content such as images, code, and text based on training data. Whether or not AI succeeds in matching the quality of a human creator is a hotly debated topic; however, what AI does offer is speed.

Fast content creation is extremely useful in construction for repeated actions such as scheduling, site planning, and project delay recovery ideation.

Speeding the Planning Process

In the site planning phase, AI models can be used to quickly assess the impact of different layouts on natural light, energy efficiency, and other factors.

“We’re actually doing a trial run with AI for office space design,” says Peters. “You take an architectural program, a list of rooms, their square footages, their adjacencies, plug it in and let the AI automatically generate different layouts for you to assess.” Pre-construction and estimating could also benefit from AI. Peters says, “It would be amazing if you could just upload a set of PDFs and have AI do the takeoff for you. We’re not there yet, but it will be possible in the future.  I predict estimating software will be incorporating some form of AI very soon.”

Evaluating Recovery Options with AI

Projects falling behind schedule is a major challenge in the building phase, which can easily start to cascade and cause additional delays if not contained as quickly and efficiently as possible. ALICE Technologies provides case studies about using their AI solutions to create recovery plans after experiencing on-site delays. In one example, a team used ALICE to optimize overtime for critical crews to recover from a 30-day project completion delay.

LLMs Are Overcoming the Labor Shortage in Construction

Trained on vast quantities of text, large language models (LLMs) can generate or summarize content, answer technical questions, and assist in knowledge transfer.

Increasing Apprenticeships with AI-Enhanced Training

The application of LLMs has huge potential for helping to train and upskill workers faster. Google is supporting the electrical training ALLIANCE (etA) as it integrates AI tools into its curriculum to boost the number of apprentices in the United States. This AI-enhanced program aims to increase the electrical workforce pipeline by 70% within the next five years. AI holds similar implications for construction labor shortages across other trades as well.

Generating Proposals with AI

To assist with answering RFPs, AI not only saves time up front in research and writing, but also points out where humans might be forgetting things. Peters says, “Once we’re 100% comfortable with governance and protecting intellectual property, we’ll be able to take some natural next steps with AI. For example, we have about two years’ worth of proposals in a database, so we can find like projects very quickly based on set of parameters. The next step would be to develop enterprise GPTs and train AI models on how we want things to be written.”

Predictive Analytics Optimizes Scheduling and Maintenance

Predictive analytics, which draws on machine learning, is already helping companies reduce equipment downtime and optimize scheduling. Using AI to schedule heavy machinery or equipment placement reduces idle time and results in fewer sequencing errors.

Replacing Repetitive Construction Tasks

AI’s immediate future in construction will see a particular focus on improving efficiency for everyday organization, communication, and planning. Construction project processes that have a lot of repetition are the obvious first places to look when implementing AI solutions. Peters says, “Scheduling can benefit from AI since there’s a lot of repetition. You could train models on past schedules, take key prompts along with the timeline, and automagically generate a schedule.”

Using Visualization to Predict Architectural Issues

What Peters is eager to see is advancement in vision capabilities. For AI to review floor plans the way an experienced architect can, it would have to be trained to know everything about drawings: what a door symbol looks like; what a dimension line is compared to a wall; how to read elevations, etc.  With that capability, AI would be able to detect issues such as stairs not being wide enough, rooms that require more doors, distances that need verification, and all the things an architect can see.

Increasing Uptime for Construction Equipment

When the Internet of Things (IoT) enabled access to sensor data, predictive maintenance largely replaced scheduled maintenance. The timing of repairs was optimized based on actual data like temperature, vibration, pressure, and usage hours. AI takes predictive maintenance further by analyzing vast amounts of IoT data so that it can detect anomalies and predict failures with high accuracy.

Agentic AI Provides Higher-Level Decision-Making

Compared to automation and AI that does what it’s been programmed to do, agentic AI can make independent decisions within guardrails. Agentic AI also adapts to its environment, and strives for continuous improvement. It’s like having a new hire who has proven themselves enough to be entrusted with less supervision and more responsibility.

Human Oversight Is Essential with AI Tools

Training, governance, and intelligent implementation are critical to ensure that the benefits of AI tools outweigh the risks. In an interview with ENR, experienced construction lawyer Peyton Aldrich said, “AI is a tool, not a replacement for human oversight. Without proper checks, you risk relying on inaccurate outputs that could lead to costly mistakes.”

Peters stresses the importance of ownership and responsibility, “We have to make sure that we’re not taking humans out of the equation, because if you’re having AI do all your work, you’re opening yourselves up to risk if people aren’t back-checking things.”

Leading stakeholders are paying close attention to actual use cases for AI in the construction industry, piloting test cases, and implementing AI where it has the best return on investment with the least risk. As construction companies thoughtfully integrate AI capabilities to complement human expertise, the industry is ready to make the most of this technological advancement.