Heat, water, and electricity. Huge amounts of energy are wasted in buildings. But with the help of AI, the Danish company Ento can map out where that waste occurs – and take action.

‍

Nurseries heated over the weekend, leaking pipes, old water heaters keeping water at boiling point—there is extensive energy waste happening in buildings, and it’s taking a toll on the climate.

– "The operation of buildings accounts for 28 percent of total CO2 emissions. And according to the International Energy Agency, energy efficiency in buildings is the single biggest factor in whether we reach our climate targets. Overall, nothing is more important than energy efficiency when it comes to climate," says Malte Frederiksen, Chief Commercial Officer at Ento, a Danish company whose mission is to reduce global energy consumption – by using AI to analyse energy data from buildings and pinpoint areas of waste. Naturally, this doesn't just benefit the climate.

– "We’ve seen examples like a toilet that was continuously running for a whole year, costing an extra 140,000 DKK in annual operating costs. Or energy waste in a single building amounting to 20,000 DKK per hour. A housing association in Aarhus recently saved 600,000 DKK per year in fines for heat waste because they identified and resolved the issue," Frederiksen explains. He adds that if a building owner has never worked on reducing energy waste, they can typically cut energy consumption by 20–30 percent across their property portfolio once they start.

‍

Many of our clients have an extremely large number of buildings and therefore lack an overview of whether they are being operated correctly.

‍

Lack of oversight

Ento was founded in 2019 and, according to Frederiksen, has doubled its customer base every year since. Today, the company monitors energy usage in 30,000 buildings across Europe, with clients including the Salling Group retail chain and 30 Danish municipalities. With a current staff of 15, Ento also serves clients in the UK – such as the supermarket chain Planet Organic – and public buildings in Milan, Italy.

– "Many of our clients manage a vast number of buildings and therefore lack an overview of whether they’re being run properly. Take a municipal kindergarten, for example: who checks if the ventilation system runs at night? Or maybe it’s a toilet in a basement of a public building that’s been leaking for years, or a burst pipe – things no one notices unless you analyse the data. Our mission is to help building owners eliminate waste in electricity, water, and heat."

‍

‍

Year-round heating

– "When it comes to electricity use, ventilation systems are the biggest culprits. These are usually controlled by sensors that monitor CO2 levels and ventilate accordingly. If the sensor malfunctions and gives inaccurate readings, the system could run 24/7 without anyone noticing. The same goes for electric water heaters keeping water constantly at boiling point.

– "We’ve also seen a ramp in a car park where heating cables were installed in the floor, meant to activate during freezing weather to prevent ice. Our system detected that the heating was running constantly all year round. Fixing it took three minutes but had cost hundreds of thousands of kroner," Frederiksen explains.

It’s also difficult to configure heating systems correctly in buildings.

– "A building’s heat is affected by so many external factors – such as solar radiation. You risk overheating when the sun shines in, because the heating system can’t account for it."

‍

We can train machine learning models to gain a unique understanding of each building’s specific needs.

‍

Heating plan and proposing solutions

This is where AI comes into play.

– "Ento offers two products: one reads consumption data to detect anomalies and identify possible causes – like incorrectly set ventilation. This relies on pattern recognition in the data. For example, if water consumption behaves a certain way, we know in 99 out of 100 cases it’s a running toilet, while a burst pipe shows a different pattern. This allows us to provide concrete recommendations.

– "The other product allows algorithms to take control of the building systems. The system knows precisely when to turn heating on or off or lower it, because we can train machine learning models to understand each building’s specific needs," he says. Take a primary school: you might want 22°C from 8 a.m. to 4 p.m. Monday to Friday, and outside those hours you only care that it doesn’t drop below 15°C, to avoid mould.

– "The algorithm can create a heating schedule based on the weather forecast and factors like outdoor temperature, solar radiation, wind direction, precipitation, etc. It also accesses calendar data, so it knows when there are public holidays and can avoid heating empty buildings."

All of this is managed via Ento’s software, which connects to existing systems in the buildings and also uses publicly available data sources. This means the algorithm receives information from sources like weather services, the Danish Building and Housing Register (BBR), and Google Maps.

‍

The algorithm can create a heating schedule that takes into account the weather forecast and factors such as outdoor temperature, solar radiation, wind direction, precipitation, and so on.

‍

Too much data for humans

Of course, running an AI system consumes energ – but according to Frederiksen, much less than what it saves. He also insists that the AI isn’t doing something humans could do instead.

– "Humans simply can’t do what our system does. In a large building portfolio, you might need 1,000 sensors, each sending hourly data points. Then you’d have to enter it all into a spreadsheet and analyse it – and no human could keep up. It’s just too much data."

‍

This interview was originally published in Danish by DM Digi.

‍