It was the kind of a deal that excites building energy nerds.

A year ago, BrainBox AI, which provides energy-efficiency analytics software to commercial real estate firms, completed an acquisition with the Swedish engineering giant ABB. The deal meant the Montreal company would now be able to remotely manage HVAC systems in more than 12,000 buildings in the U.S.

Though vast, BrainBox AI’s new portfolio itself is unremarkable: mostly big box retail stores. But what the company’s AI-based software does is optimize the energy consumption of each one, thus driving down both heating and cooling costs, as well as associated carbon emissions. That’s what got the building nerds excited — small savings add up. Across the portfolio, its software for HVAC systems yields energy reductions of up to 25 per cent and cuts greenhouse gas emissions by as much as 40 per cent. “For small retailers,” says Nicholas Bossé, BrainBox AI’s chief energy transition officer, “these benefits are very important because it goes directly into their bottom line.”

Despite such results, Bossé, who is a veteran of the energy conservation world, knows how tough it is to convince property managers to adopt new technology. “The thinking,” he says, “is that ‘if it’s not broken, you don’t need to fix it.’”

The wrinkle, as he points out, is that these quotidian systems are often much more broken — which is to say inefficient and therefore costly — than their owners realize.

The need to optimize our building stock

The workaday world of HVAC technology is filled with unsung contraptions like boilers, chillers and air handling systems. But while Class A office buildings are now kitted out with smart sensors and other state-of-the-art energy management technology, the same can’t be said of much of North America’s commercial real estate, which includes everything from drafty warehouses to the quick-and-dirty strip mall developments that account for so much suburban retail.

All those buildings emit an enormous amount of carbon. When you factor in the impact of gas furnaces and such carbon-intensive materials as concrete, the construction and building sector, including millions of apartments, houses and other structures, generates a whopping 37 per cent of all carbon emissions globally.

To make things even more challenging, the transition to low-carbon electricity will coincide with steep increases in demand due to the proliferation of electronic devices, electric vehicles and heat pumps. Utilities and local energy distributors, in turn, need to undertake a massive reorientation away from the sector’s traditional reliance on power plants and transmission corridors to a much more decentralized approach that incorporates smaller-scale renewables, battery storage, smart grids and energy efficiency technologies.

State-of-the-art energy management software plays a critical role in this shift, largely due to the fact that it’s able to reduce carbon emissions without huge capital outlays. (Software, after all, is much easier to implement than switching to hydrogen.) 

To help expedite this transformation, a growing number of Canadian firms have developed AI-based tools. With algorithms capable of accurately predicting energy demand and optimizing electricity consumption, this kind of automation allows building owners to drive down their own energy costs, incorporate renewable electricity sources and push consumption to less expensive periods. In other words, these smart solutions make it easier to decarbonize our built environment.  

The power of prediction

When it comes to managing complex building energy systems, one of the biggest benefits of AI is that it can eliminate subjectivity and human error, says Sue Talusan, director of the Impact Accelerator at MaRS. “It removes some of the human elements from decision-making that are not necessarily helping and allows us to be anticipatory, not reactive.”

For instance, BrainBox AI’s software can make slight adjustments to heating levels based on anticipated traffic in a building to optimize energy consumption. It connects with the HVAC management system in a building, absorbs the patterns of energy use and then feeds all that data into its algorithm to predict future demand. The software also incorporates such details as weather and time-of-day utility rates and building occupancy, then sends instructions to the control system to make subtle shifts in temperature settings, which helps save on energy.

Toronto-basedPeak Power helps building managers reduce energy consumption and meet environmental goals through different means. It deploys large-scale batteries and rooftop solar installations that can store inexpensive power for use during high-cost periods. “The main driver of the AI is around forecasting and operations,” says CEODerek Lim Soo. “It’s able to predict the most expensive and dirtiest hours of highest electricity use.” The firm’s software draws on a range of data, such as weather satellite images, building info and shifting power rates to maximize the client’s reliance on its own power sources while minimizing higher cost and dirtier electricity from the grid.

These kinds of platforms, adds Talusan, can provide automation and energy optimization across a portfolio of buildings. “You can now monitor and influence the behaviour of multiple assets with this type of technology in a proactively automated and programmed way,” she says. “You don’t need a human to look after each asset individually.” For chains with portfolios of dozens or hundreds of outlets, this kind of automation makes it much more straightforward for them to hit their energy efficiency goals.

Such systems can be highly scalable, says Bossé. Case in point: BrainBox AI recently expanded a four-store pilot for Sleep Country Canada to 214 outlets. The approach involved deploying “cloud-connected and AI-enabled” thermostats in each outlet, thus allowing BrainBox AI’s technology to operate across the entire portfolio. Sleep Country’s electricity and gas bills for these sites have dropped 24 per cent and 22 per cent respectively. “My role,” Bossé says, “is to use that data and not only to have the building be, in and of itself, more efficient, but also to contribute to the energy transition.”

Finding a good flow

AI has become a useful tool for a growing number of energy efficiency ventures focused on finding new ways of reducing building-related carbon. For example, as EVs become more popular, the electricity demands facing both local distribution grids and property owners will also become more complicated. But these vehicles hold a lot of potential — as an energy source. Numerous startups are working with charging networks and manufacturers to deploy bidirectional charging in offices or apartment buildings, enabling EV owners to both charge their batteries or allow them to be used as backup power in the event of a blackout.

SWTCH Energy, a Toronto-based firm that operates about 10,000 EV chargers in apartment buildings, offices and retail outlets across North America, uses AI to predict demand and avoid operational glitches in its equipment. “Reliability is a big challenge with EV chargers,” says CEOCarter Li. The firm’s network is constantly tracking its charging fleet, and uses performance analytics to predict the likelihood of service interruptions so the firm can carry out preventive maintenance before a given charger goes out of service.

Until the advent of EVs, policy and technology experts viewed transportation — and building — related carbon emissions separately. EVs, as well as rapidly developing charging technologies (including those monitoring networks to ensure reliability) have blurred those lines, and not just in the parking garages of high rises. Companies that rely on fleets transitioning to EVs, such as distribution centres, bus depots and manufacturing facilities, all face similar questions — not just about increasing on-site electricity consumption, but also how that power should be managed.

Devashish Paul, founder and CEO of Ottawa-basedBluWave-ai, works with utilities, transit operators and logistics companies to optimize their fleet charging requirements. He mentions a company in Qatar that operates a fleet of 3,000 electric buses that was looking to fast-charge them safely in Qatar’s heat. “They’re killing the batteries of these buses and they really don’t know how fast to charge them, but they have a route schedule to meet.”

The firm’s AI-based smart grid can monitor large numbers of EVs as well as renewable power sources, such as rooftop solar. In Qatar, Paul explains, “we needed to predict how much energy is needed for the bus fleet to go do its routes, and then from the grid connection, basically figure out exactly how much energy is needed for each bus and how slowly we can charge it without overheating the battery.”

With all these technologies, the goal is to leverage the power of AI to drive energy efficiency and ultimately reduce the building sector’s contribution to the climate crisis. But for all their power, these platforms depend on the tough questions that building owners and developers are prepared to ask themselves.

For climate-conscious firms like Sleep Country Canada, they’ve realized the search for a low-carbon future may be found on flat roofs that can hold solar arrays, as well as in cloud-based AI- systems that control HVAC gear. “The real estate sector in general is conservative,” observes BrainBox AI’s Bossé. “They know that they have to evolve and to become more digital. We try to do as much as possible with the data we gather. The core technology is really energy efficiency, so we’re saving energy and dollars for the operator.”

Smarter — and more people-friendly — buildings

While many so-called smart building platforms now incorporate AI to drive down energy costs and reduce carbon, these tools have other more human-oriented applications, too. For instance, the federal government’s office at 25 St. Clair East, also known as theArthur Meighen Building, has gone through a $277-million overhaul since 2019. The building is meant to serve as a showcase for cutting-edge applications of AI in both energy and accessibility.

Among the features: an 82 per cent reduction in greenhouse gas emissions from pre-renovation levels and an AI-driven building management system. But, as Sue Talusan, director of the Impact Accelerator at MaRS, points out, the project involved extensive investments in both reconciliation — through collaborations with Indigenous designers — and accessibility. The project features extensive sensors and state-of-the-art digital mapping tools that enable visitors and employees with physical and visual impairments or hearing loss to seamlessly navigate the space. These smart applications can help improve buildings — for everyone.

BrainBox AI is one of eight companies in Mission from MaRS: Public Procurement, a special initiative that’s working to make it easier for communities to adopt climate solutions. John Lorinc writes about technology for MaRS. Torstar, the parent company of the Toronto Star, has partnered with MaRS to highlight innovation in Canadian companies.