For a long time, much of facilities management has followed a fairly straightforward cycle: something happens, a team responds, and the problem gets fixed.
That model isn’t going away. Buildings will always produce surprises, equipment will fail, and facilities professionals will still need the experience to respond when something unexpected happens. But today, buildings are producing far more information that can help teams spot potential problems earlier.
HVAC and building management systems, occupancy sensors, energy monitoring, work orders, equipment performance, and workplace platforms all generate data about how a building is operating. Artificial intelligence offers a way to make better use of that information, finding patterns across large volumes of data that would be difficult for a person or team to identify manually.
The opportunity for facilities management is to gradually shift more of the work from responding to what has already happened toward understanding what is likely to need attention next.
Finding problems earlier
Predictive maintenance itself isn’t new. Industrial and manufacturing environments have been monitoring equipment performance for years because unplanned downtime can be enormously disruptive and expensive. What is changing is the amount of building data now available and the ability to analyze it at scale.
Instead of relying only on a fixed maintenance schedule or waiting for equipment to fail, facilities teams can use operational data to identify changes in performance. A system drawing more energy than usual, equipment behaving differently from its historical pattern, or environmental conditions beginning to drift can provide an early indication that something deserves attention.
IBM’s 2026 overview of AI in facilities management describes this as one of the core applications of AI in FM: combining real-time and historical sensor data to identify equipment likely to fail and allowing maintenance to take place before a breakdown.
I think it’s important, though, not to overstate what the technology can do. An algorithm isn’t necessarily going to tell a facilities manager with certainty that a chiller will fail on a particular day. What it can do is identify an unusual pattern and give an experienced professional a reason to investigate.
Used that way, AI isn’t a substitute for facilities expertise. It helps people apply that expertise more efficiently.
In industrial environments, the stakes can be higher
The potential becomes especially interesting when we look beyond the traditional office.
In manufacturing, logistics, laboratories, and other operationally intensive environments, the facility is closely connected to the ability of the business to function. A building system failure may not simply make occupants uncomfortable. Depending on the operation, it can interrupt production, affect product quality, create safety concerns, or contribute to significant downtime.
That changes the economics of facilities management. If data helps a team identify deteriorating performance earlier, maintenance can potentially be scheduled before a problem becomes an emergency. Parts can be ordered in advance, and work can be coordinated around production schedules rather than interrupting them.
These applications aren’t necessarily the most visible examples of AI, but they may prove to be some of the most useful. In a manufacturing environment in particular, a building that simply operates reliably can have significant business value.
Better data can change how we use buildings
Maintenance is only one part of the opportunity. Corporate real estate teams are also trying to understand how buildings and spaces are actually being used, something that has become more complicated as work patterns have changed and organizations reconsider what they need from their portfolios.
Occupancy data can show that a building is underused. Energy data can show what it costs to operate. Work orders may reveal recurring trouble spots, while building systems provide another layer of information about asset performance. Looking across those signals together can give a CRE team a very different understanding of a property than any single dataset could provide.
That makes the quality of the underlying data increasingly important. JLL’s 2026 Global Occupancy Planning Benchmark Report found that improving space-data accuracy has risen to the second-highest CRE priority, behind portfolio optimization. The report makes a straightforward point that is relevant well beyond occupancy planning: AI-powered tools are only as useful as the data feeding them.
The same research also points to growing attention around technical spaces, including laboratories, manufacturing and distribution facilities, warehouses, and data centers. These are often among the most operationally complex and expensive spaces in a portfolio, making better information about how they are being used especially valuable.
From facilities data to investment decisions
Once organizations have a clearer view of building performance, the implications extend beyond day-to-day operations.
Every CRE team has to decide where to put limited capital. Equipment age and maintenance history have traditionally been part of those decisions, along with inspections, budgets, and the experience of facilities professionals. Better operational data adds another layer by showing how an asset is actually performing over time.
That becomes particularly useful across a large portfolio. Someone who knows one building intimately may recognize that a particular system is beginning to behave differently. Finding similar patterns across hundreds of buildings and thousands of pieces of equipment is much harder.
AI and analytics can help identify where something looks unusual so teams can focus their attention there. Over time, that information can also help organizations make more informed decisions about whether to repair equipment, replace it, invest in a building, or reconsider the role of an asset in the portfolio.
This broader shift is already underway. JLL’s 2025 Global State of Facilities Management Report found that 28% of organizations had embedded AI solutions into their FM operations, rising to 46% among organizations with more than 100,000 employees. JLL identifies predictive maintenance, work-order management, and performance tracking among the areas where AI can produce meaningful operational returns.
AI can’t solve a bad data problem
There is a temptation with any new technology to jump ahead to the end state: connect the systems, apply AI, and make the building predictive.
The reality is usually messier. Many organizations are managing buildings with systems installed at different times, from different manufacturers, using different standards. Data may be incomplete or difficult to access, and one property may have sophisticated building management systems while another has relatively limited digital capabilities.
Those differences matter because AI doesn’t make poor or fragmented data disappear. If organizations don’t have a clear understanding of what data they have, how reliable it is, and whether their systems can work together, the output will be limited as well.
JLL’s facilities management research highlights both data quality and integration with legacy systems as important challenges to successful AI deployment. That suggests some organizations may get more immediate value from improving their data foundations than from rushing to adopt the newest AI platform.
For CRE and facilities teams, the starting point should be the problem they are trying to solve. From there, they can determine what information is needed, whether they have it, and where technology can genuinely improve the decision-making process.
The role of facilities professionals will change with the technology
Greater automation naturally raises questions about what it means for facilities jobs. In the near term, I see this less as a story about replacing people and more as one about changing where they spend their time.
If technology gets better at monitoring systems, sorting through information, and flagging potential issues, facilities professionals can spend less time searching for problems and more time evaluating them. An unusual data point still needs context. A recommended repair has to be weighed against budgets, business operations, occupant needs, safety, equipment life, and other considerations that may not be apparent from the data alone.
That puts a premium on both traditional facilities expertise and newer skills. Understanding buildings and equipment remains fundamental, but data literacy and the ability to work across increasingly connected systems will become more valuable.
Technology can help teams spot things they might otherwise have missed, but facilities professionals still need the experience and context to determine what that information means and what, if anything, they should do about it.
A more informed approach to building operations
Not every facility needs the same level of technology. A sophisticated manufacturing plant operating around the clock will have very different requirements from a lightly occupied office, and the business case for predictive maintenance will look different in each.
What these facilities have in common is the opportunity to make decisions with better information. For some organizations, that will mean advanced predictive maintenance. For others, the more useful first step may be connecting isolated systems, improving asset data, or developing a clearer picture of occupancy and energy use. What matters is whether the technology helps facilities teams reduce unnecessary downtime, use resources more effectively, extend the useful life of assets, and support the people and operations inside their buildings.
Facilities management will never be completely predictive. Buildings are too complex, and unexpected problems will always happen. But as the quality of building data improves and the tools for interpreting it become more capable, teams have a growing opportunity to identify issues earlier and make decisions with more context.
That’s a much more practical measure of AI’s value in facilities management than whether a building can be called “AI-powered.”
