Facilities Management Technology

Technological innovations are driving rapid change in facilities managers’ work, underscoring the need for cross-functional collaboration, data analytics and effective risk management, speakers said at the NMFT Remix conference in Orlando in late October. 

Generative AI

ChatGPT holds promise for facilities managers, said Rafi Dowla, co-founder and CEO of PremiseHQ, a Toronto-based maker of systems for analyzing numerous facility management functions. The generative artificial intelligence tool can help enhance operational efficiency, reduce costs, boost tenant satisfaction and provide other competitive advantages, Dowla said. His team has collaborated with clients to strategically introduce ChatGPT to their internal teams, with a focus on striking a balance between minimizing risk and garnering a sizable return on investment, he said. 

“Thinking about a data pipeline strategy is critical before deploying ChatGPT directly in your system,” Dowla said. “How you process your data and integrate with ChatGPT is … the most important part of the strategy. Behind the user interface, you need to work with an [AI] expert who can help you architect that in a way that is scalable and portable.”

One example Dowla gave is the potential for ChatGPT to offer insights into indoor air quality improvements. His team is conducting a study in concert with a Canadian university to deploy sensors for identifying hot and cold spots in buildings. “We’re using ChatGPT to analyze that data and give feedback to the buildings in terms of where we need to put systems in place to address these hot and cold spots,” he said.

But the use of generative AI faces adoption barriers such as testing and optimization, Dowla said, as well as issues related to system compatibility, data security, privacy, employee training and tenant education.

Sensors and the Internet of Things

Applications of the Internet of Things in facilities management include integrated workplace management systems that use sensors to measure occupancy, air quality and the efficiency of HVAC systems, speakers said. Sensors can detect motion; sensors placed under office desks can determine occupancy based on how much body temperature is generated by the person seated there; and sensors deployed in air quality monitoring devices can monitor indoor and ambient environments, said Peter Costanzo, director of workplace technology solutions firm ROI Consulting Group. 

The proliferation of air quality sensors needs to be matched with better hardware that can go into air filtration systems as well as more staff training on servicing those filtration systems, according to Kevin Delahunt, senior technical advisor at BGE Indoor Air Quality Solutions. “We shouldn’t have to put [in] a gas casket and such. But we do, as we’re trying to make the best of a bad situation. We need to look at the hardware as we go up in efficiency,” Delahunt said.

Costanzo emphasized the evolving role of thermal sensors and digital twins — virtual representations of a facility — in smart building functions amid a shift from wired to wireless technologies. “There’s also the concept of a gateway or a hive or hub, which I would call a proprietary network. What’s nice about some of the gateways … is that they’ll work at different frequencies [which means] you can cut down battery usage [and thus] cut down the amount of battery that’s needed,” Costanzo explained. Additional technologies like lidar and indoor positioning systems are effective in addressing challenges of efficient space utilization, wayfinding, managing and tracking assets in large facilities, and occupancy monitoring, Costanzo said. 

Data insights and cybersecurity

Data insights are crucial for gaining access to capital and maximizing the service life of equipment, said Sergio Pages, client relations manager at StructureTec Group, a consulting and construction services firm that specializes in restoring building envelopes, roofs, pavement and concrete structures.“Data drives our decisions. We want to have a plan of attack, and we need data that helps us make those decisions [in terms of] where to deploy funding when it comes,” Pages said during a session on creating better buildings. 

Pages advises facilities managers to tap into those data insights to optimize the efficiency of newer equipment. “If I put in $100,000 for a new asset, I can get 10 years out of it. Or I can put in $150,000 and get 20 years out of it. When you get the funding, take into consideration the service life costs of what you’re going to be investing that money on,” Pages said. 

Facilities managers can harness technologies like real-time energy optimization and monitoring platforms — based on fault detection and diagnostics — to monitor and reduce energy consumption, said Ronak Shah, energy analyst at InSite. 

Stephen Stalnaker, vice president and general manager at the Florida division of Albireo Energy,  said he has been receiving questions about how buildings are measuring their carbon footprint, waste output, water output and energy usage for government reporting purposes. “So we’re working with facilities managers to produce that data,” Stalnaker said at a session on the role of data analytics in facilities management. 

The increased technological functionality of buildings has also spawned a host of issues related to cybersecurity, Stalnaker pointed out. Although outsourcing cybersecurity has become the norm, particularly in multi-acre facilities spanning thousands of buildings, in-house technology experts are important for security, Stalnaker said. “Cybersecurity and IT support are people you would want to have in your staff. In data centers, it’s very critical. So, IT support and cybersecurity are typically almost always in-house, and you’re going to have to have someone onsite,” Stalnaker said.

Building portfolio owners can identify building decarbonization opportunities more quickly by applying machine learning, artificial intelligence and physics-based modeling, according to a Nov. 29 report released by McKinsey.

By using data from satellites, geospatial analytics, regulations, labor and equipment costs, and assessing heating and cooling systems, insulation levels and the feasibility of solar or geothermal power, algorithms can analyze and propose solutions for building portfolios to reach net-zero emissions. 

With this new approach, financially optimized plans, which account for the regulatory environment and the building’s unique characteristics and leasing structure, can be developed for an entire portfolio within weeks, experts said in the report. 

Given that buildings are responsible for 40% of global combustion-related emissions, it is essential to cut 50% of direct building emissions and 60% of indirect emissions by 2030 to achieve net-zero carbon building stock by 2050, the McKinsey experts said. Traditional methods of decarbonization, involving physical energy audits and building-by-building net-zero strategies, are considered laborious and expensive, McKinsey said. Additionally, a lack of centralized inventory and standardization contributes to the perception that decarbonizing buildings is unprofitable.

Compared to traditional energy audits and net-zero studies, AI-driven approaches provide a more than 100-fold increase in the pace and scale of decarbonization planning, eliminating the need to rely on vague building archetypes, the report stated. 

The authors highlighted the potential for an AI-based approach to generate neutral to positive returns on investment for real estate portfolios, assuming the absence of factors like incremental future regulations, carbon pricing and green premiums on rent or property valuation.Implementing energy efficiency and electrification measures for each building, while optimizing renewable energy procurement at the portfolio level, enables building owners and occupants to recoup their investment by realizing energy savings, optimizing their capital costs and avoiding regulatory penalties, the paper emphasized. 

Features of an optimal building decarbonization plan

McKinsey emphasized that achieving the most effective building decarbonization plans involves seven components that can be optimized through the use of AI and machine learning methods:

1. Efficient net-zero planning: Owners can ensure coordinated, comprehensive plans for their entire portfolio through joint procurement and strategic sequencing, unlike traditional decarbonization plans that often target select buildings based on emissions or existing regulations. 
2. Asset-specific plans: Tailored plans that consider aspects like building layout and type of insulation are needed for cost-effective decarbonization. Each building requires a unique strategy that considers its starting point, local conditions and asset specifics like tenant composition and lease structures.
3. Complete pathways to net-zero: This involves avoiding partial plans that compromise long-term outcomes. Companies must take comprehensive, forward-looking decisions as short-term strategies risk increased costs and overlook synergies like insulation measures that affect future HVAC requirements.  
4. Integrated Scope 1 and Scope 2 plans: Disjointed approaches toward energy efficiency and electrification hinder efficiency, the report said. The failure to exploit interdependencies can lead to slower and costlier renewable power procurement.
5. Actionable steps: Building plans must provide precise instructions for facilities managers and enable easy communication between vendors and facilities management teams to ensure swift execution.
6. Quantification: Plans must be specific enough to offer detailed insights for financial planning, encompassing net-zero goals, capital investment challenges, operating costs, potential debt and the allocation of costs and benefits between building owners and tenants, so that leaders can understand the exact costs of achieving net-zero emissions.
7. Net-zero-oriented decision-making: Owners and operators can integrate decarbonization plans into organizational operations by adjusting processes, incentives and governance structures. This involves updating capital plans, budgeting for low-emission systems and incorporating decarbonization analyses into new asset acquisitions. 

Decarbonization challenges related to scaling supply chains to meet new demand and training skilled workers to deploy retrofits and undertake other electrification efforts also impact the industry, the report said. 

Fortunately, adopting an AI-backed, full-life-cycle approach to decarbonization simplifies planning, accelerates processes, and reduces costs, enabling significant progress in addressing building-related emissions, McKinsey said.

Honeywell launched Advance Control for Buildings, a building management system that can use a building’s existing wiring “to give building managers more control over the efficiency of their buildings, ultimately helping improve the occupant experience while advancing energy management goals,” it said in a press release Tuesday. 

The new BMS platform seeks to leverage Honeywell’s recent partnership with NXP Semiconductors to elevate machine learning and autonomous decision-making for energy management and its integration of Analog Devices’ technology for greater digital connectivity and advanced signal conversion across applications, it said. 

“Today, buildings are facing mounting pressures around safety, operational efficiency and energy efficiency. By pushing the boundaries of technological advancement, Advance Control not only helps building managers optimize their building operations to meet these pressures, but also champions innovation,” Steve Kenny, vice president and general manager of Honeywell Building Management Systems, said in the release. 

According to the UN Environment Programme, buildings account for 37% of global carbon dioxide emissions, leading to growing pressures to identify and reduce energy use, Honeywell said. It noted that a lack of skilled labor also makes it challenging for building owners to hire technicians for buildings’ upkeep and modernization. As a result, building owners and operators are turning to digital building and energy management tools. Schneider Electric estimates its products can slash up to 42% of operational carbon emissions in commercial office buildings, for example. 

Automation and machine learning drive the platform’s efficient operating system, which is equipped with integrated cybersecurity and technology, to deliver accelerated network speeds, Honeywell said. Automation is able to break down operational barriers to give building managers more control over the efficiency of their buildings, ultimately helping improve occupant experience and advancing energy management goals, Kenny said in the release. 

Last October, Honeywell announced plans to realign its business segments with megatrends like industrial automation and the energy transition to boost its organic sales growth and deploy more inorganic capital. The company said the implementation would begin in the first quarter of 2024. 

Its new advanced control system aligns with its growing focus on its building technologies segment, which CFO Greg Lewis called “a bright spot for the business” due to its growing profit margin despite soft demand for security products and a challenging operating environment. 

Advance Control joins other building automation and security-focused solutions Honeywell has launched in the past year, including Forge for Buildings, which the company introduced to optimize air quality and reduce energy consumption, and a new integration with Quantinuum that intends to bolster the security of smart utility meters.

Inventory management firm MarginPoint unveiled a new computerized maintenance management system platform aimed at commercial real estate, government, university, construction and multifamily managers, according to a Dec. 18 news release. The platform, designed to streamline operations for facilities management teams, trade contractors and distributors, provides tools for third party coordination and inventory, asset and work order management, according to MarginPoint. 

By centralizing inventory and supply chain management, alongside asset data, the company can provide “a complete ecosystem that enables efficient, profitable management of your facilities and the maintenance of them," MarginPoint CEO Vince Sheeran said in the release.

Due to the high costs of parts and labor, and costly repairs associated with the failure of critical building assets, facilities managers are increasingly turning to computerized maintenance management systems, or CMMS, to gain insights into asset health, streamline maintenance operations and budget capital expenses and upgrades. 

Providers and solutions aimed at accommodating that need include Hexagon, eMaint, ServiceChannel, Accruent, Planon, Cityworks and JLL, through its Corrigo platform. That competitive landscape has also pushed multiple firms to raise funds in pursuit of scaling their products and establishing dominance in a market that seeks cost-effective maintenance solutions. 

CMMS firms Limble and MaintainX, for instance, have raised significant funds this year on the heels of impressive reported revenue growth. Limble raised a $58 million series B in June, following a 130% increase in revenue over the previous 12 months, while MaintainX secured $50 million in funding earlier in December, after clocking a 13-fold increase in revenue since 2021. 

MarginPoint says the key features of its new CMMS platform include automated supply and replenishment, asset management, work order management, preventative maintenance, API-enabled integrations and AI-driven visualizations and insights. The platform can help manage and automatically replenish inventory to individual properties, cut maintenance costs, improve maintenance timing and increase expense ratios and tenant satisfaction by allowing facilities managers to complete more jobs in a day, MarginPoint said. 

By automating the scheduling of regular maintenance tasks and tracking the health of assets on its platform, it can help facilities managers reduce maintenance costs, prevent downtime and extend asset lifespans, the release noted.

JLL has developed a large language model for the commercial real estate industry, according to a release. The generative AI model, JLL GPT, is powered by in-house data and external commercial real estate sources, and will be used to provide insights to the firm’s global workforce and clients. 

The product is the next step in JLL’s development and deployment of AI, which has included implementing the technology to improve building efficiencies, generate 3D leasing visualizations and calculate sustainability risks. 

JLL says the generative AI solution follows the growing emergence of artificial intelligence in the facilities management space, as the industry increasingly adopts the technology to manage building controls and drive portfolio-level decision-making, amid significant labor concerns, high energy costs and rising tenant demands.

Developed by in-house technology division JLL Technologies, JLL GPT will be deployed to the company’s 103,000-plus person workforce and is set to provide consulting experts with a quicker turnaround on comprehensive workplace planning solutions for clients. 

As facility operators and building owners continue to look to technology to create operational efficiencies, JLL plans to roll out made-to-order client solutions later this year. 

“The most interesting thing that I’ve seen in my technology lifecycle is the emergence of GPT, large language models and generative, pre-trained transformative models,” Sharad Rastogi, CEO of JLL Technologies’ work dynamics technology group, told Facilities Dive. “We had the internet, we had mobile and now we have GPT. It’s going to be a very seminal transformation that is going to impact real estate quite dramatically.”

“Using generative technology, we can fundamentally improve the efficiency, the productivity, the insides and make buildings run a lot better,” Rastogi added. 

Additionally, the company is increasingly leveraging the technology to drive investment leads and noted that one in five of JLL Capital Markets opportunities during the first quarter of 2023 was enabled by AI capabilities on its platform. 

Rastogi added that brokers will be able to use the technology to automatically generate marketing material, analyze leases and query databases to make more efficient decisions. 

JLL noted this technology is the result of a data-first philosophy, which has provided it with the clean, quality-assured and strictly governed commercial real estate data that is needed to extend into generative AI technology. Rastogi said that while the company has already put in considerable effort to these ends, there is still room for improvement. 

“There's always more to do, right? We're always enhancing that [data], and I think that's going to be a journey. We're finding more and more data sources, and I think as the buildings get smarter, the number of endpoints with generative data are growing and are going to increase dramatically,” Rastogi said. “The better your machine learning models are, the better your forecasting is going to be and the better the insights you're going to get. This is a self-fulfilling prophecy and a self-reinforcing cycle”.

He also believes that the real estate industry as a whole is in the early stages of generative AI development and will see rapid growth in the next few years.

In a survey conducted by ASIS International, only 8% of professionals reported that they had not experienced any access control failures in the past six months.

Serious access control incidents are few, however, respondents said, with 71% reporting no more than five per year. The most prevalent problem was tailgating or piggybacking, mentioned by 61% of respondents. Tailgating is when an unauthorized person follows an unaware user into a restricted area; piggybacking can refer to such access being provided with the authorized user’s knowledge.  

Despite those findings, access control systems — the combination of technology and security policies used to permit or deny access to a location — are highly effective in preventing unauthorized physical incidents, the report states.

The report by the global association of security professionals surveyed over 1,000 frontline security, management, director and C-suite level professionals, from those operating at a single facility or campus to those managing security at multinational organizations with a variety of facilities in multiple countries. It found that 93% of organizations believe access control is an essential piece of their risk management or security plan. 

When security consultants were asked which access control concerns or failures are most common, 34% said that problems arose from poor implementation of technology or from choosing technology that is not correct for the situation; 24% said failure resulted from over-prioritizing employee convenience; and 13% attributed failures to technology that is insecure or can be easily hacked. 

The report identifies a trend toward adoption of more secure technologies. Low-frequency proximity cards, while previously widespread, have been phased out due to their vulnerability to replication or spoofing, and more secure technologies have taken their place, it states. Whereas 51% of respondents reported using low-proximity cards in 2019, only 22% said they use them in this year’s survey. Instead, 50% of organizations now employ what the report calls highly secure credentials, such as mobile devices, biometrics and multifactor authentication, with 35% using medium-level security measures such as RFID proximity cards and smart RFID cards. Professionals who deploy higher-level security systems expressed more confidence in their overall access control systems: 70% of those with highly secure technology agree that their system is effective, compared with 58% of those with less secure technology. 

The report also highlighted the crucial role of access control in visitor management, with 39% of respondents still using manual systems, such as spreadsheets or pen and paper, to keep track of temporary credentials issued to visitors. 

Organizations also commonly turn to video cameras to complement access control systems with surveillance, with 85% and 84% of respondents using inside and outside surveillance, respectively. When asked how they ensure access control systems have not been breached, 76% of security professionals said they actively monitor the video surveillance.

Carrier Global Corp. has agreed to help develop and commercialize AirJoule, a dehumidification and cooling technology for HVAC systems from Montana Technologies that Carrier called “transformational.”

The AirJoule technology provides evaporative cooling by harvesting thermal energy and water from air, in some cases without refrigerants. It has the potential to cut electricity consumption by roughly 75% compared with conventional HVAC systems, Montana Technologies says. 

Carrier has also conditionally pledged a $10 million growth equity investment in the commercialization of AirJoule. The agreement potentially gives Carrier the exclusive right to commercialize AirJoule across HVAC systems in the Americas for three years, according to a separate Jan. 8 news release from Montana. 

Montana Technologies’ partnership with Carrier and the latter’s investment in it demonstrate both companies’ commitment to “reducing energy consumption, cost and carbon footprint,” Montana CEO Matt Jore said in the news releases. The agreement follows Montana’s partnerships with chemical producer BASF and lithium-ion electric battery manufacturer Contemporary Amperex Technology Co. 

Last June, Montana announced it will combine with Power & Digital Infrastructure Acquisition II Corp., or XPDB, and list on the NASDAQ pending completion of the merger. That transaction is expected to be finalized in the first quarter of this year. Carrier will receive the right to nominate a member to the board of directors of the merged company, the release said. 

“At Carrier, we are committed to investing in both sustainable and disruptive solutions that will better our planet for generations to come,” Ajay Agrawal, chief strategy officer and senior vice president of business development and global services at Carrier, said in the release. “Our partnership with Montana Technologies, and our future collaboration to commercialize its AIrJoule technology, marks another step forward in this commitment.”

Schneider Electric is incorporating Microsoft Azure OpenAI into several products and services aimed at leveraging the capabilities of generative artificial intelligence to create text, code and other forms of content for customers and internal operations.

By using generative AI, the company hopes to streamline time-consuming tasks, optimize resource allocation and boost the efficiency of operational processes, Schneider Electric said in a Nov. 13 press release.

The company used a large language model to add a conversational AI tool to Resource Advisor, its energy and sustainability data hub. The Resource Advisor Copilot will aid facilities managers where swift aggregation of data on energy consumption and related resources is crucial, Sreedhar Sistu, vice president of AI at Schneider Electric, said in an interview.

With their potential to accelerate operational efficiency, cut costs and improve tenant satisfaction, generative AI tools can be useful for facilities managers, according to Rafi Dowla, co-founder and CEO of PremiseHQ, which automates data management for various facilities management functions across commercial and industrial properties. 

Large language models are expected to spur improvements in efficiency, productivity and the overall performance of buildings, Sharad Rastogi, CEO of JLL Technologies’ work dynamics technology group, told Facilities Dive earlier this year upon the rollout of JLL GPT, a large language model aimed at providing commercial building operators more efficient access to building controls, leasing, utilization and portfolio optimization insights.

However, despite gaining some traction in facilities management, tools like ChatGPT, often encounter adoption barriers related to data security and privacy, Dowla said. 

Schneider Electric’s Sistu echoed that sentiment. “The biggest concern with tools like ChatGPT is that it is public and it scrapes off the internet, and then whatever you feed into it also becomes public. So, there are risks associated with it, and concerns have been raised [about the lack of] security and privacy,” he said.  

Resource Advisor Copilot aims to equip Resource Advisor customers with enhanced data analysis, visualization, decision support and performance optimization, strengthening their ability to process intricate industry knowledge, according to the company’s news release. Unlike traditional methods that involve searching for and synthesizing data, Resource Advisor Copilot lightens the workload for facilities managers and provides the “right context for information” in a more natural and streamlined way, Sistu said. A significant portion of this information exists in textual form that requires interpretation, understanding and processing, he noted. 

Through its OpenAI integration, Schneider Electric is using two internal conversational AI assistants — Jo-Chat GPT, which provides a secure platform for Schneider employees, and Finance Advisor, which helps financial analysts locate precise information that can inform their decision-making in a consistent and timely manner. Sistu suggests that internal conversational assistants could be a template for professionals in facilities management and related functions who are looking to incorporate secure and private generative AI platforms within their teams. 

The company is also exploring Github Copilot, a tool provided by Microsoft, for creating offers and conducting operations. “Internally, it allows us to be more efficient in writing code and developing applications,” Sistu explained. He anticipates that this type of emerging technology will be beneficial in developing specialized code for control systems in buildings. The intention is to leverage such technology for domain-specific language used in building control systems which may have “unique logic and syntax” not readily available in the public domain, he noted. 

Cushman & Wakefield announced a new artificial intelligence strategy aimed at creating a new industry standard for commercial real estate from proprietary data, strategic technology partnerships, clients and employees. As part of the AI+ strategy, it will embed artificial intelligence across its commercial real estate transaction lifecycle to boost productivity for its operations and clients and help its brokers, services and research professionals support client decision-making, the company said in a Nov. 20 news release.

The announcement marks the growing integration of AI into commercial real estate as service firms attempt to streamline operations and increase operational efficiencies. 

Generative AI investments are on the rise across the country, according to a Lucidworks survey, which found that 92% of U.S. companies it surveyed plan to increase AI investments in the next 12 months. This trend has increasingly taken hold in the commercial real estate service and facilities management industries, as JLL, CBRE and now Cushman & Wakefield deploy strategies to optimize their operations. 

JLL released JLL GPT earlier this year, which uses in-house data and external commercial real estate sources to provide insights to the firm’s global workforce and clients. CBRE announced that more than 20,000 client sites worldwide are using its AI-based Nexus platform to improve efficiency and reliability across over 1 billion square feet of property. 

Cushman & Wakefield says its AI strategy is designed to deliver digital transformation at scale. Deployment will increase access to its enterprise expertise and insights and the speed at which they can be deployed to the market, the firm said. 

By “seamlessly integrating” data, people, processes and technology,  “we can magnify the impact of secure and well-planned digital transformations for our clients and colleagues,” Salumeh Companieh, chief information and data officer, said in the release. 

The firm began using AI in 2018, it said, with a focus on aligning its business, data and operations. Cushman says to date, AI has enabled an 80% material reduction in its operational cycle time and a reduction of client supply chain costs by optimizing its proprietary supply chain network. 

Technologies the company has already implemented include C&W Atlas Mapping, a self-service, interactive mapping program it launched in September, and “marketing copilot” Jasper. 

“Through AI+ … Cushman & Wakefield can bring clients the ‘CRE platform of the future’ today through an approach that also provides a nimble, scalable, and secure operating model for addressing forthcoming needs of owners, occupiers, and our firm,” Andrew McDonald, global president and chief operating officer, said in the release. 

Cushman & Wakefield did not immediately respond to a request for comment.

Technology’s shining star this year continues to be generative AI, and commercial real estate companies refuse to be left out after playing catch-up in digital maturity compared to other industries. 

Commercial real estate leaders expect AI and generative AI to have the greatest technological impact on the industry over the next three years, according to a JLL survey of 1,006 decision makers conducted by Meridian West. The only technology that rivals AI's position are solutions that support sustainability.

Sandeep Davé, chief digital and technology officer at CBRE Group, and his technology teams have evaluated AI integrations across the lifecycle of real estate, from the minute an asset is identified, through the decision to invest and into the management phase.  

“We’ve got AI interventions across the board,” Davé told CIO Dive. “The way we think about any technology investment, including AI, is how is it aligned with our strategy and what is the business problem that we are trying to solve.”

CBRE wanted a model that could work with real estate data and documents. The solution was a sandboxed innovation environment. 

“We’ve created a self-service AI playground, which we are adding about 150 users on a weekly basis on pure word of mouth,” Davé said. “It’s where our teams can come in and query real estate data, have the models point to real estate documents, query the documents and translate them into multiple languages.”

What's in the sandbox

Commercial real estate companies are under significant pressure from clients to adapt to their evolving digital needs, from employees to craft better user experiences and from economic headwinds. 

Interest rates have undergone the "sharpest rise in nearly 40 years,” CBRE said in its October earnings call presentation. The jump in rates pushed an expected market rebound to the second half of next year at the earliest, according to CBRE CEO Robert Sulentic, speaking during the company’s Q3 earnings call. But there are still pockets of opportunity, Sulentic said. 

CBRE serves more than 95% of Fortune 100 companies and has more than 500 offices spread out across more than 100 countries, according to its website. 

The company has deployed an AI-enabled smart facilities management capability to more than 20,000 sites covering one billion square feet. CBRE also uses AI to manage millions of dollars worth of supply chain analytics, according to Davé. 

Yet, for CBRE, the risks generative AI pose are clear. In order to mitigate some of the fear around exposing proprietary or sensitive information, the company blocked access to public generative AI capabilities. 

The concern comes from the idea that employees could input confidential information in a query that would then be later used in a generated response to a separate user. It’s a real and well documented issue for enterprises that want to harness the power of generative AI without exposing sensitive data. 

“In today’s environment, I see a lot of hype and fear of missing out,” Davé said. “Whereas our approach has been very consistent and aligned with our strategy so we get really outsized, meaningful returns.”

With generative AI entering the mainstream tech stack conversation, CBRE has leveraged its previous experience working with transformer models to put a specific focus on productivity benefits within job families. 

“For example, software engineers in my organization use GitHub Copilot to help them with writing code and generating test scripts, almost a pair programming kind of a capability, but there are other roles we’re looking at [like] marketing support,” Davé said. 

The company has “lots” of small pilots underway, but Davé’s teams have found such significant benefit from the coding assistant that the tool is rolled out broadly, according to the tech chief. 

While developers in the company mainly leverage GitHub’s Copilot coding assistant, the rest of the company uses the AI playground to generate content, chat and ask questions. The company has around 3,000 unique users hopping on the platform on a monthly basis. 

“Rather than worrying about what specific use case I’m enabling, we actually enable a range of self-service capabilities,” Davé said. 

Even as the company adds to the platform, it’s cautiously monitoring costs, according to Davé. CBRE is tracking analytics to measure platform usage, costs and efficiency gains to help strategize the continued roll out and the rest of its AI investments. 

“We’ve got to ensure that we are getting good returns for the investment,” Davé said, stressing strategic alignment and clarity. 

“What we do is to drive the growth or efficiency of the business, and we’ve reinforced that for our AI program, but it’s also just become part of our culture,” said Davé.