A word on… good data
Terry Rose, HVAC Service Director at Integral UK, discusses the importance of foundational data in creating effective smart buildings and explains why, despite the excitement about new AI capabilities for integration, there is still a strong need to get the basics right first.
Like every other sector, management is eager to adopt artificial intelligence (AI). A recent global study by JLL found that almost 87% of teams have started piloting AI, or at least plan to this year, up from just 5% two years ago. The sharp increase is driven largely by C-suite mandates. If it’s the job of business leaders to look at the horizon, their view is currently dominated by AI.
Unfortunately, the same study found that only 5% of organisations running these AI pilots reported achieving all their programme goals. This gap between ambition and delivery is a familiar story to those responsible for technology and operations. The pressure is falling on technology teams, facilities managers and engineers to make AI projects deliver results, often without the data quality or infrastructure needed to support them.
The issue is not the technology itself but the information it relies on. There’s a certain irony in that AI has replaced data as the fashionable word, but the former can’t function properly without the latter. Clean, reliable data gives organisations an accurate picture of how their buildings perform, providing the basis for a plan and the reasoning behind it. Before any organisation invests in AI, it needs to be clear about what it’s trying to achieve. In that sense,
the first question should always be, ‘why?’. What problem does the technology solve? What outcome does it create? How does it support business or operational goals? For facilities teams, this might mean improving reliability, reducing manual reporting or cutting energy use. Once the purpose is defined, the next step is assessing whether the right data exists to support it. Without that foundation, any AI initiative will struggle to deliver measurable results.
Why good data matters for HVAC
HVAC systems show this clearly. They already generate large volumes of data through sensors, meters and building management systems, but the value of that information depends on accuracy and consistency. When data is structured and complete, it supports better decisions at every level. Engineers can check that plant operates within design limits, identify drift in performance and plan maintenance before faults cause downtime.
Reliable data also supports the transition to reliable condition based maintenance. Instead of fixed service intervals, engineering teams can use performance trends to decide when work is needed. This avoids unnecessary visits, focuses resources where they are most useful and extends the life of assets over the long-term. This kind of visibility helps shift from decision-making based on guesswork to an environment of measurable performance.
When AI adds real value
Once the data foundation is in place, AI can support plant management in practical ways. It can analyse performance trends, detect anomalies and suggest control adjustments that improve efficiency. When combined with other building or operational information, such as weather forecasts or occupancy, it can help maintain comfort and reduce energy use.
Again, these benefits depend on the quality of the underlying information. This makes the verification and validation of data more important, because AI can only work with information that has been checked and confirmed as accurate. If sensors are miscalibrated, meters faulty or asset data incomplete, for example, the analysis will be wrong.
The JLL research suggests then that organisations making real progress with AI are those that invest first in data and system integration. Building services teams should take the same approach. AI can improve speed and analysis, but it can’t compensate for missing or poor-quality data.
While the current focus on AI has drawn attention to technology in property and engineering, it risks skipping over the basics. In HVAC and building management, good performance still depends on accurate measurement and dependable records. AI can be a powerful part of the process once those are in place, but it must never be the destination.
