These days, energy sector headlines involving artificial intelligence are focused on how utilities can meet the significant power demand of data centers, but perhaps lost in the shuffle is how AI is being applied to the building sector to boost energy efficiency and reduce carbon emissions.
While AI has emerged as a technology to enhance productivity and improve life quality, “its role in building energy efficiency and carbon emission reduction has not been systematically studied,” researchers at the Lawrence Berkeley National Laboratory noted in a study published in 2024, “Potential of artificial intelligence in reducing energy and carbon emissions of commercial buildings at scale.”
The study evaluated artificial intelligence’s potential in the building sector, focusing on medium office buildings in the United States. A methodology was developed to assess and quantify potential emissions reductions.
Key areas identified were equipment, occupancy influence, control and operation, and design and construction.
Six scenarios were used to estimate energy and emissions savings across representative climate zones.
“Here we show that artificial intelligence could reduce cost premiums, enhancing high energy efficiency and net zero building penetration,” the study’s authors said.
Adopting artificial intelligence could reduce energy consumption and carbon emissions by approximately 8% to 19% in 2050, the study said. “Combining with energy policy and low-carbon power generation could approximately reduce energy consumption by 40% and carbon emissions by 90% compared to business-as-usual scenarios in 2050.”
In a Q&A with Public Power Current, Nora Wang Esram, CEO of the New Buildings Institute, was asked to detail the ways in which utilities can leverage AI to boost energy efficiency in buildings in their service territories.
“First, it’s important to recognize that AI is a very broad term. In buildings, we already have a range of tools for modeling and analytics, advanced controls and automation, and load forecasting and reduction,” she said.
“Some of these use machine learning techniques, but they are not the type of AI most people think of today -- like large language models -- even though some companies rebrand these tools as AI. These existing technologies have significant potential to improve building efficiency, lower peak energy use, and free up grid capacity to support electricity growth, but they remain underutilized. Before investing in entirely new AI technologies for buildings, we should first look at what we already have and what has been proven to work,” Esram noted.
That said, “these tools haven’t been fully adopted for reasons such as cost, complexity, or lack of interoperability. This is where new AI capabilities could make a difference. New AI tools based on large language models can recognize patterns, interpret images, communicate in natural language, and handle complex or tedious tasks through simple prompts. In doing so, they could make existing building tools more accessible, more affordable, and better connected across different data systems,” she noted.
Esram also said that there are multiple knowledge gaps when it comes to awareness of the potential uses of AI when applied to building operations.
“First, some utilities, regulators, and service providers associate AI with futuristic technology, and they may not fully recognize that machine learning is already being applied effectively in buildings and the grid, as well as the new capabilities promoted by tech companies,” she said.
“Second, AI is only as good as the data it’s trained on. High-quality building data that are publicly available are still limited, which constrains AI’s potential in building operations until better data structures and sources are in place. There’s a real opportunity to educate utilities on where to invest resources and how to create pathways to use these new AI capabilities to improve building efficiency and strengthen grid resilience.”
New Buildings Institute is a nonprofit organization working to advance scalable and practical approaches that equitably eliminate emissions from the built environment. “Our efforts help keep energy costs affordable, cut emissions that are fueling climate change, and deliver on improved health, safety, and resiliency for everyone,” it says on its website.
Esram was also asked to provide examples of how AI is being used in an effective way to boost energy efficiency in buildings.
“As I mentioned, machine learning is already being used in advanced building management systems to predict occupancy patterns and optimize HVAC schedules accordingly. These systems can also incorporate weather forecasts and grid signals, such as emissions or pricing, to adjust lighting and HVAC settings with minimal disruption to building operations,” she said.
Regarding newer AI capabilities, “one common use case is using AI to review and process documents, which can speed up permitting processes. Some organizations are also exploring the combination of multiple data sources—such as satellite imagery, building energy use data, and publicly available building characteristics—to identify opportunities for building upgrades,” Esram said.
“Another emerging application is using AI’s language capabilities to provide training, help building operators diagnose issues and create dashboards, and engage building occupants. The cost-effectiveness of these new tools and approaches is still being evaluated.”
She also discussed how “digital twins” can be paired with AI to optimize building operations.
“Digital twins can be expensive to create because they require detailed building data, specialized software, and ongoing updates to stay accurate. However, as tools become more automated and AI-assisted, costs are coming down, making them more accessible for typical buildings,” Esram said.
“The more digital twins we deploy, the greater the aggregated benefits we can achieve for the grid. AI can also make digital twins much easier for building operators to use. It enables operators to interact in plain language, automatically highlight patterns or issues, and provide guidance or recommendations based on the building model. This is critical to achieving expected energy reductions -- high-performance buildings that are mismanaged won’t deliver results,” she noted.
“In short, AI helps realize the impact of these technologies by making them more accessible and lowering the entry point for handling complex tasks.”
AI also has the potential to expand digital twins beyond energy management, such as enhancing safety and security (e.g., in schools, assisted-living facilities, etc.) and coordinating other essential building functions (e.g., in food services, airport), Esram went on to say.
“Essentially, AI can turn digital twins into a multi-dimensional tool that supports overall building performance and business operations. Imagine if walls could talk.”
Schneider Electric Sees Early Success in Commercial Real Estate Buildings Leveraging AI
Schneider Electric has observed early success in commercial real estate buildings leveraging AI, dependent on the building’s lifecycle including:
• Design and construct more sustainable buildings: AI-driven tools allow engineers to consider sustainability metrics from the earliest design stages, including elements like low-carbon materials, optimizing shade, and maximizing solar utilization. These intelligent AI-driven designs can contribute to more aesthetically pleasing and environmentally responsible buildings.
• Operate and maintain to reduce a building’s environmental impact: HVAC systems are often the primary energy consumers in buildings. AI uses real-time data on occupancy, weather patterns, and a building’s thermal properties to optimize HVAC settings, achieving an optimal balance between energy efficiency and occupant comfort. AI-powered predictive maintenance can help minimize the risks of costly repairs.
• Full asset lifecycle to enhance the tenant experience: AI transforms facility management by providing actionable insights from analyzing data points across the building. Now, facility managers can make proactive, data-driven decisions that strengthen asset and operations performance. AI is being leveraged to optimize HVAC systems, manage microgrids, and predict maintenance needs, helping enhance the full tenant experience.
Company Uses Advanced Deep Learning Algorithms to Predict Building Energy Needs
BrainBox AI, which was acquired by Trane in early 2025, offers a suite of platforms including ARIA (Artificial Responsive Intelligent Assistant), which “works to transform the daily routines of those who work to keep our facilities running smoothly -- simplifying operations, enhancing decision-making, and optimizing building performance,” as BrainBox AI notes on its website.
“Acting as a personal building ally, ARIA enables facilities managers and building operators to command their building's operations, from RTUs to chillers, boilers, and more, through voice or text -- transforming building data into precise insights and strategic actions in real-time,” BrainBox AI said.
The company also offers advanced AI for HVAC optimization.
With over 7,000 locations already connected to BrainBox AI’s Envoy EMS solution, Dollar Tree sought to retrofit an additional 600 stores with a focus on energy optimization. To achieve this, they piloted BrainBox AI’s autonomous AI HVAC Optimization solution, covering 6.6 million square feet across diverse geographic regions -- from New England to California. This approach allowed Dollar Tree to assess the solution’s scalability and adaptability to varying climates while maintaining uninterrupted operations.
BrainBox AI has posted a case study on the Dollar Tree project on its website.
Powered by the data extracted from Dollar Tree’s BMS, combined with external data sets such as weather forecasts, BrainBox AI’s autonomous HVAC optimization solution effectively transformed energy usage patterns across Dollar Tree’s BrainBox AI-enabled stores. This resulted in reduced energy consumption, portfolio-wide visibility, and greater data-driven decision-making, reducing unnecessary technician dispatches and saving thousands per avoided trip, BrainBox AI said.
In just one year, 600 stores throughout the U.S. achieved electricity savings of 7,980,916 kWh, helping Dollar Tree meet its interim emissions targets without large-scale equipment overhauls.
The reduction, driven by BrainBox AI’s advanced algorithms, translated to a total cost savings of $1,028,159 by optimizing equipment runtimes. Additionally, the Dollar Tree team noticed less overall work orders raised at these locations.
The pilot’s results have driven further deployment of BrainBox AI’s AI HVAC Optimization solution across more than 2,000 more Dollar Tree stores.
Study Reveals Commercial Building Managers Plan to Increase the Use of AI
In February, Honeywell released the findings of its AI in Buildings study, which revealed that 84% of commercial building decision makers planned to increase their use of AI in the next year to help them improve security, streamline energy management and integrate predictive maintenance.
In this study of U.S. building managers and decision makers with more than 250 building occupants, Honeywell found that across property types, respondents were increasingly using AI to help improve process efficiency, productivity and operations.
“However, since most respondents (92%) reported challenges in hiring skilled, tech-savvy individuals, a greater opportunity remains ahead for building operations to tap into AI's capabilities to enhance employee training, augment their current workforce and ultimately help upskill labor to support the sector's rapidly changing needs,” Honeywell said.
Honeywell in June 2025 announced the launch of Honeywell Connected Solutions, “an AI-powered platform that integrates critical building software and technologies into a single interface to help enable more efficient operations.”
The platform's early adopters – Verizon Communications Inc. and Vanderbilt University – have already begun using the solution in their buildings.
With Connected Solutions, which is built on Honeywell Forge, building operators can manage Honeywell software, systems and devices through one integrated interface.
When a building or campus is fully connected, users gain comprehensive data and real-time visibility into how critical systems are operating alongside actionable insights on how to troubleshoot challenges that may arise, the company said.
Users, which can include facility managers, multi-site operators and integrators, can connect to the platform quickly with an AI-enabled installation process that is completed in hours, dramatically reducing labor time, costs and disruption, compared to traditional building management systems, it noted.
The solution also addresses the key issues facing buildings today through capabilities including:
• Advanced encryption to help safeguard against cyberthreats,
• Remote monitoring and diagnostics that help to reduce labor time and cost,
• Predictive maintenance prompts to spot and address issues before they escalate and
• Energy-management solutions supporting decarbonization efforts.
“After integrating Honeywell's Connected Solutions into some of its global footprint, Verizon is using the platform to help predict critical building and system issues before they become serious and costly,” Honeywell said.
Vanderbilt University is also using Connected Solutions across its campus, with nearly 10% of its buildings already adopting the technology. The university aims to enhance building system efficiency, reduce energy consumption and optimize the user experience, particularly in older facilities.
Siemens Offers AI Solution Through Building X
In a report, Siemens says that the building domain is trailing behind other industries in adopting and utilizing AI, noting three significant challenges:
1. Collecting, organizing, and managing data
2. Upscaling the deployment of smart building solutions; and
3. Dealing with the complexity and uniqueness of buildings
Siemens says that it addresses these challenges with its Building X, “a holistic, open platform of data-driven applications and connectivity solutions for buildings during the operations phase.” The platform “is based on extendable business services and a common data model that provides a single source of truth for a digital building.”
