AI IN CONSTRUCTION

The Three Boardrooms Shaping the Conversation

These days, whether you are in a project update meeting, at an industry event, or even catching up with friends, AI eventually finds its way into the conversation.

And after having many of these discussions, I came to the conclusion that most AI conversations in construction usually happen in one of three boardrooms.

The Three Boardrooms

Boardroom 1: Momentum Without Understanding

In the first boardroom, the conversation is driven entirely by momentum rather than understanding. The pressure to adopt AI is already there, even when people are not entirely sure what it actually means.

I once had an executive tell me, “I don’t even fully understand AI yet, but I know we need it because everybody else is moving in that direction.”

And honestly, that probably summarises more AI conversations than people would like to admit.

Boardroom 2: Technology Without Context

In the second boardroom, the dynamic shifts completely.

A meeting has been scheduled, and the room is full of people attending because they were asked to, not because they genuinely want to be there. A technology provider walks through a lengthy presentation filled with technical jargon. The slides look impressive. Everyone nods politely. But somewhere along the way, the conversation completely loses touch with the reality of delivering a project.

I recall a senior executive quietly pulling out a beard comb mid-presentation and slowly combing his beard while staring into absolute oblivion as predictive analytics and machine learning models filled the boardroom screen. At that point, it was obvious the conversation had lost the audience.

Boardroom 3: Where the Conversation Finally Becomes Real

Then there is the third boardroom.

A technology provider understands construction, and a customer understands the actual predicament they are trying to resolve.

I recall one discussion where the project director interrupted the AI presentation five minutes in and simply asked, “Can this help us detect critical path slippage before it starts impacting the project?”

And suddenly, the conversation changed. It was no longer about buzzwords. The focus shifted toward real problems, real decisions, and real delivery impact.

That is when the conversation finally becomes real. Unfortunately, most organisations today are still meeting in the first two boardrooms.

AI Is Not the Beginning of Technology in Construction

There is clearly a lot of excitement around AI. Some people are intimidated by it, while others are still trying to determine whether it is truly transformational or simply the latest industry trend.

But the bigger issue is that many conversations around AI start in the middle instead of at the beginning.

People hear terms like predictive analytics, generative AI, machine learning, and agentic AI before anyone explains the fundamentals. As a result, AI starts sounding far more complicated than it actually needs to be.

Construction itself is not new to technology.

The industry has used ERP systems, scheduling platforms, dashboards, reporting tools, analytics engines, and automation workflows for years. Most projects are already overloaded with systems, dashboards, spreadsheets, and disconnected platforms.

AI did not suddenly introduce analytics into construction. We already had analytics, dashboards, and automation.

What changed is speed, scale, and pattern recognition.

Experience and Pattern Recognition

Anyone who has spent enough time in construction naturally starts recognising patterns.

When procurement slips, delays usually follow. When staffing drops too low, productivity suffers weeks later. When approvals slow down, schedules become compressed downstream.

That is experience.

AI follows the same principles, except instead of learning from a handful of projects over many years, it can process millions of data points across thousands of activities simultaneously.

And that changes how decisions get made.

Most traditional systems explained what had already gone wrong. AI creates the possibility of identifying risks, trends, and disruptions earlier — before project teams are fully in recovery mode.

Understanding the Different Types of AI

Even within the AI space itself, the terminology has become confusing.

Conversations quickly jump between predictive analytics, generative AI, machine learning, and agentic AI, often as though they all mean the same thing.

The reality is they do not.

Generative AI

Generative AI is designed to create content.

It can summarise meetings, generate reports, organise project information, draft RFIs, answer questions, and help teams retrieve information faster.

Predictive AI

Predictive AI focuses on forecasting.

It analyses historical and live project information to identify patterns and warn teams about risks before they fully impact the project.

Agentic AI

Agentic AI moves beyond simply generating or predicting information and starts taking action.

Construction is not new to automation. The industry has operated for years on rule-based logic such as, “If X happens, do Y.”

For example:

“If a material delivery is delayed by more than three days, send an alert.”

Useful — but limited.

The system follows instructions, but it does not understand the bigger picture.

Agentic AI is different because it can evaluate multiple conditions simultaneously and adjust based on what is happening around it.

Take a real scenario. A manpower shortage develops on a critical activity. A rule-based system sends an alert. Someone reads it, adds it to the next progress report, and by the time a decision gets made, the schedule has already slipped by two weeks.

An agentic AI system handling the same situation would identify the shortage, trace which downstream activities are at risk, calculate the likely schedule impact, notify the relevant teams, and surface recovery options — all before the next morning report.

The alert does not sit in someone’s inbox waiting for a meeting. The response starts happening.

In simple terms:

• Generative AI helps create

• Predictive AI helps forecast

• Agentic AI helps initiate action

Why This Matters in Construction

Construction is one of the most operationally complex sectors there is.

Capital projects involve thousands of moving parts across schedules, procurement, labour, safety, logistics, subcontractors, approvals, regulations, compliance, and cost controls — all operating within highly dynamic environments.

Most project teams are already drowning in information. The challenge is no longer access to information. It is making sense of it quickly enough to make accurate and up-to-date decisions.

AI brings value by strengthening teams, not replacing them — giving earlier visibility into patterns, trends, and potential risks.

In scheduling, AI can detect signals that often lead to delays before the schedule fully slips. In productivity management, it can identify trends linked to workforce distribution, sequencing, or resource constraints before performance visibly drops. In safety and supply chain management, it can help teams identify growing risks before they begin affecting execution on site.

Most systems used to explain the past. AI gives teams a better chance of seeing what may happen next. At the same time, AI is not a shortcut around operational discipline.

Adding AI on top of a broken workflow does not fix the workflow. It simply produces faster, more detailed evidence that something is broken.

Disconnected systems, poor data discipline, and unclear accountability do not disappear because a new platform was introduced. In many cases, AI simply makes those gaps impossible to ignore.

Technology does not fix broken operations. People, process, accountability, and execution still matter.

The Real Question

Before the engineering, procurement, and construction sector races toward the next AI platform, tool, or trend, there is a more uncomfortable question worth asking first.

How many projects right now are carrying risks that nobody has spotted yet — not because the data does not exist, but because nobody had the time or the tools to connect the dots?

That is the real conversation.

Not whether AI is transformational or overhyped, but whether teams understand it well enough to know where it genuinely helps — and where it does not.

Successful adoption of AI is not about moving faster toward technology.

It is about clarity, readiness, and understanding how technology can genuinely improve decision-making, operational visibility, and project delivery.