Why we need to go back to basics when it comes to AI

March 9, 2026

As in most industries, artificial intelligence continues to dominate the conversation in facilities management. Predictive maintenance, automated scheduling, and AI-powered asset insights are all being presented as the next big leap for the field.

According to Ivan Morley, director of growth at Thermatic, the industry is at risk of skipping an unglamorous but important step. Before FM organizations can talk about AI, they need to get their asset data right.

>

From an engineering perspective, we look at AI the same way we look at other systems. In other words, if the input is wrong, the output is also wrong. This is a simple principle, but one that is often ignored. In FM, AI promises to automate decision-making, identify patterns, and reduce risk. However, all of these results depend on complete, accurate, and structured asset information. Without that foundation, AI will not produce intelligence but simply automate bad decisions.

solid foundation

Despite years of digital transformation efforts, many organizations still struggle with the basics of asset management. In practice, this means an asset ledger that is incomplete, outdated, or riddled with duplicates. Systems often don’t integrate seamlessly, and inconsistent input from engineers compromises data quality. If this is your starting point, layering AI on top will not solve anything, it will only magnify the problem.

In such situations, the AI ​​could create a dangerous illusion of control. Dashboards look impressive and reports look sophisticated, but the intelligence underneath is fundamentally flawed. In facilities management, where capital investment, compliance, and operational risk are closely intertwined, that false sense of trust can quickly become costly.

fix the foundation

Our approach is intentionally pragmatic. Asset data is not treated as a management byproduct of maintenance, but as a real operational tool. All assets are QR coded and managed through the CAFM system. Individual codes are scanned on site to ensure that work is recorded against the correct plant every time. Asset history is updated in real-time as each scheduled maintenance visit, reactive repair, or intervention occurs. This means our asset list is a working tool rather than a static spreadsheet.

The discipline behind consistent asset tagging, structured engineer notes, and real-time updates is often underestimated, but it’s what enables meaningful insights. Over time, you will be able to compile a detailed maintenance history that supports lifecycle analysis, sparing strategies, risk profiling, and performance benchmarking.

Daily tasks will also be improved. Engineers arrive in the field with accurate information, improving first-time fix rates and reducing unnecessary follow-up visits. Data quality is not treated as an abstract benefit. It directly affects service delivery.

Where AI brings real value

Once structured and reliable asset data is available, AI will truly begin to gain ground. With clean lifecycle data and failure history, trends and patterns begin to emerge that are difficult for humans to discover at scale. This could include assets deteriorating sooner than expected, sites increasing reactive spending, or equipment that is technically compliant but operationally unreliable. These are things that AI is very good at highlighting.

When used properly, AI can interrogate large amounts of reliable data and provide early warning of risks. Support more informed lifecycle planning, tighter budgeting, and more timely investment decisions. This helps teams move from being reactive to proactive intervention. What it offers is no shortcut and is entirely dependent on the infrastructure that supports it.

Risks of neglecting the basics

Attempting to implement AI on poor data is not only ineffective, but also risky. With incomplete or inaccurate asset registers, AI models will draw incorrect conclusions at a rate that actually makes them appear confident. This is when organizations begin to make bad capital decisions, which can have serious consequences.

The real power comes when AI supports, rather than replaces, the judgment of skilled engineers. Only then can the industry move forward. In a field that is increasingly focused on the potential of intelligent systems, we need to stay grounded. Sound engineering practices remain the foundation. However, if you get the basics right, AI is possible. If you skip them, it’s mostly noise.

Photo: Image of a glowing microchip with the word “AI” written in the center of a digital circuit board.

Article written by Dave Mapps | Published March 9, 2026

share

---

---

Related articles

---