How AI is influencing facilities management

For Brett Ellis, the regional facilities and asset manager at construction and mining multinational Komatsu, AI is the difference between days and weeks.

Komatsu has been using drones to map construction sites for a number of years. The information is used to determine slope angles, digging depth and cut and fill ratios. That data assists the operator to carry out work more efficiently.

Now artificial intelligence can identify defects from the drone data and rank them in order of repair priority – high, medium and low. It can provide repair cost estimates, how soon it can be done as well as the entire costing for the whole building, Ellis said at a recent JLL webinar on developments for facilities managers in the APAC region.

“I can get something done in three days, whereas before it used to take three weeks,” said Ellis. “If I do that every year, I have a baseline and I can review that. And I can get that data to talk to other systems.”

JLL research shows that technology is the common thread in how facilities managers will face a future of more work with less staff, changing asset usage, greater cost scrutiny and a need to focus on sustainability. The rapidly evolving category of generative AI is increasingly becoming an integral part of how to make all the pieces of the puzzle fit. McKinsey research estimates that generative AI could add up to $4.4 trillion of value annually.

Technology can provide facility managers with prescriptive data and insights by combining data integration, advanced analytics, and AI-driven recommendations. The collection of comprehensive data from various sources within the facility portfolio would include IoT sensors, building management systems, maintenance records, occupancy data, energy consumption data, and more, says Ellis.

And while preventive maintenance continues to dominate as a strategy among facilities managers concerned about scaling back on costs in asset management, “predictive” technology is the new buzzword, experts and practitioners say, capturing information about assets that will help drive down costs.

The use of IoT sensors and devices to monitor vibration, temperature, pressure and flow have been on the scene for quite a while now, but the introduction of AI means “we can gather even more usable data,” Ellis said.

“Machine learning algorithms analyse data from sensors and historical maintenance records to predict when maintenance is needed. AI can identify patterns of equipment behaviour that precede failures,” he said. “Those algorithms are chewing away at the data in the background and will tell you, ‘Hey you need to do something now, we noticed some trends through the sensors, we think now’s the right time’.”

And into the future, developments in technology will empower facilities managers with a more integrated approach. Savings in one area could be re-allocated to another at the swipe of a screen. Data from various sources will be able to communicate with other datasets.

“I’m definitely looking for data to work harder and technology to work harder than it did before,” says Ellis.

Kevin Janus, JLL Technologies technology program director, says the “maintenance journey” is moving on rapidly from preventive maintenance, which was state-of-the-art up to five years ago, and itself an evolution from corrective maintenance (“it’s broken, I need to fix it”).

“A lot of companies are beginning to look at this more holistically,” Janus said on a webinar to discuss the findings of JLL’s survey of facilities managers. “Predictive maintenance is really the next step.”

Janus says, “the democratization of information gathering is on us,” and as technology becomes commonplace in facilities management, the cost of acquiring systems will fall. The price of sensors is already coming down, he said.

But there are also ways in which to adapt existing systems. “Look at what you have – people were sold building management systems in the 1980s and 1990s. There are ways to take the old and the new and create a data set from it.”

A November 2022 research paper by analysts at the Civil Engineering Research and Innovation for Sustainability Institute in Lisbon, Portugal, found that throughout the entire operation and maintenance stage, facility management teams collect and process data from different sources, often needing to be adequately considered when making future decisions.

This data could feed statistical models based on AI, improving decision-making, they said.

With the emergence of intelligent buildings, which embed most spaces with smart objects, building information modelling, or BIM, provides builders with new opportunities to upgrade these buildings at lower costs and shorter project duration, allowing information exchange between the various stakeholders involved. Conventional FM practices need to incorporate the integrated approach of intelligent management, which is embedded in information and functional integration, according to the authors.

It’s now a lot easier to input data sets to get results and AI now provides a lot more information than previously, Janus says. AI is actually machine learning as well and there are systems in place, such as IBM’s Watson supercomputer which can take trend data that is too much for the average person to look at and provide trend analysis based on parameters that the facility manager might be interested in, such as seasonal factors.

But, Janus cautions, artificial intelligence “is only as good as the programming and the thought that goes in it. So, it takes human intelligence to generate artificial intelligence. That’s my cautionary tale for the day.”