The Dangerous Power of Data in Construction

Watch the episode on YouTube right here!

Talking about data in construction might sound like a technicality reserved for IT departments. But when you look at how projects are actually run—the “headless chicken” scenarios on-site, the reactive decision-making, and the billions lost to inefficiency—you realize that data isn’t just a tool; it is the foundation of the industry’s future.

We explored this topic in our latest podcast episode featuring Sara, Vice President of Data Analytics and Decision Intelligence at COWI. As an engineer by training, Sara has seen firsthand how the industry struggles to bridge the gap between traditional engineering and the rapid evolution of data science.

In the following article, we’ll discuss:

1. The “Headless Chicken” Scenario
2. Standardization is Control
3. The Data and AI Trap: A Double-Edged Sword
4. Where Will the “Connectors” Come From?
5. Reflection: What are you leaving behind?

The “Headless Chicken” Scenario

The construction industry is famously reactive. Decisions are often made under extreme time pressure simply to keep works moving, regardless of whether those decisions are optimal. Elia describes this in the podcast episode as “running around like a headless chicken”.

By relying solely on gut feeling or limited experience, we subject projects to human bias and unforeseen risks. Data and evidence-rooted facts provide a “robust way of putting substance” to daily decisions, ensuring they are based on what has actually happened rather than what we assume.

Standardization is Control

If you don’t define what you need at the start of your project, you leave room for misalignment and conflict. Sara points to ISO 19650 as a cornerstone for this transformation. It forces stakeholders to think about the end purpose of information before a single brick is laid.

• Data Economy: More data is not always better; it can actually lower quality and waste limited resources. 
• Metadata is Context: Numbers without context—like costs without currency or dates—are rendered useless for long-term benchmarking. 
• Purpose-Defined Quality: Data that is “good” for predicting maintenance on a bridge might be “bad” for calculating its immediate carbon footprint.
  

While efficiency and risk management might be the primary drivers, this structured approach is also what allows us to accurately track sustainability and carbon lifecycle analysis.

The Data and AI Trap: A Double-Edged Sword

There is a lot of hype surrounding Artificial Intelligence, but Sara is clear about the hierarchy of value.

“To have good AI, you absolutely need good data… If the input quality is poor, your output will definitely be poor. There’s no other way around it.”

While data is often hailed as the “new oil” for the construction industry. The “dangerous power” of data lies in the gap between having the technology and understanding the responsibilities that come with it.

Here are the primary dangers we must navigate:

• The Garbage In, Garbage Out Trap: Using advanced AI models on poor-quality data is one of the biggest risks in the industry. If the input quality is low, the output is guaranteed to be poor, leading to flawed decisions that sit on “shaky ground”. 

• Security and Public Safety Risks: As we move toward data-rich BIM models for infrastructure like international airports or metro stations, data becomes a “precious treasure” for those with malicious intent. Sharing models on unverified platforms without understanding where data is stored creates massive security vulnerabilities.
  

• The Accountability Gap: As AI begins to make safety-related decisions on-site, the industry faces a looming legal crisis. If an AI system makes a mistake that leads to an accident, the lines of responsibility are currently blurred, making regulatory frameworks like the EU AI Act essential for future protection.
  
• Misguided “Data Democratization”: There is a dangerous trend of providing everyone access to all data without limits. Without strict permissions and adherence to GDPR, sensitive intellectual property and personal data are at constant risk of exposure.

The AI landscape is moving fast, with thousands of platforms emerging and platform storage policies shifting within months. To remain in control, the construction industry cannot treat digital transformation as an afterthought or wait for slow-moving regulations to set the rules. Instead, we must anchor this rapid change in robust internal governance—establishing clear policies, access permissions, and ethical responsibilities—to ensure we use this power to create value rather than damage. 

Where Will the “Connectors” Come From?

To manage this “dangerous power,” the industry needs more than just data scientists; it needs Connectors. These are hybrid professionals who act as translators between the world of algorithms and the reality of a construction site.

Where do they come from?

• Engineers evolving into data roles: Like Sara, many engineers are “pivoting” by adding data literacy to their technical training. 
• Micro-learning: The younger generations are expected to pivot 4–6 times in their careers, relying on continuous, purpose-driven learning rather than a single university degree. 
• Cross-disciplinary degrees: We are seeing more hybrid profiles, such as professionals combining environmental science or law with data analytics. 
• A New Education: Perhaps an entire new academic degree in construction data science is what it takes to supply these Connectors?
  

It is key that these individuals understand the business value. They should know that a 0.001% increase in model accuracy isn’t worth it if the model becomes unexplainable or loses its practical application. It should be usable for construction projects.

Reflection: What are you leaving behind?

The international construction industry is at a signaling point. We have the tools (right now) to save between 10% and 15% in construction costs through proper AI and data usage – according to Autodesk. But this transformation doesn’t start with an algorithm; it starts with a plan.

As you go through your workday, consider the data you are interacting with right now. Every spreadsheet, every site photo, and every procurement quote is a data point. Are you capturing it? Do you have a plan to use it a year from now? If the answer is no, you are leaving the industry’s most valuable asset—information—on the table.