Does AI actually cut heating costs? Two years of testing revealed the answer

Samuel Gibbs was doubtful that handing control of his heating to technology would make the house cheaper to run while still staying warm. In the end, the outcome proved more positive than he expected.

Smart thermostats and AI systems are often marketed as a way to trim heating bills, with manufacturers touting average savings of around 25% or roughly £180 a year. But those claims come with plenty of caveats, and shifting energy prices make it hard to know what is really driving any change.

To get a clearer answer, I compared two years of my own heating habits: a strict manual approach with a simple schedule versus an AI-managed system. I used my 1930s solid-brick house, which has a recently insulated loft conversion, as the test case.

Warming up

At the start of the test period, my setup consisted of a 10-year-old Worcester Bosch combi gas boiler running at a medium setting, a Hive thermostat in the hallway with its smart functions switched off, and 13 radiators fitted with standard thermostatic radiator valves, or TRVs.

To keep energy use down, I used a basic timetable that switched the heating on for a couple of hours in the morning and again in the evening. I closed off the TRVs in rooms we were not using and adjusted others by hand when they did not need to be heated. I also switched the heating off manually whenever I left the house.

The routine was tedious and sometimes uncomfortable. Other people in the family would glare at me while wrapped in blankets in the living room, or accuse me of being a traitor when I refused another request to turn the heating up. Even so, the approach kept winter bills low and more money in my pocket.

After allowing for hot water use by comparing summer readings when the heating was not in use, the manual setup used 14,316.7 kilowatt hours of gas over 213 days. That was around 10% to 15% lower than in earlier years, when I had been less strict.

The switch

Later, I replaced the Hive thermostat with a full Tado X system linked to the existing combi boiler. The setup included 13 smart radiator thermostats in place of the standard TRVs, a wireless thermostat in each of the two largest rooms, and a programmer wired directly to the boiler.

Rather than using one central thermostat and a single temperature for the whole house, the heating is controlled room by room. Each smart TRV can call for heat from the boiler and then regulate hot water flow into the radiator. When a room does not need heat, the TRV shuts the radiator down, helping the boiler run more efficiently because it needs to burn less gas to keep the water hot.

If a boiler supports the OpenTherm protocol, it can also be told to lower its flow temperature and further reduce energy use. My boiler does not support that function.

The Tado smartphone app manages the whole system and makes it easy to copy schedules across rooms or fine-tune each room separately.

The equipment is not cheap. The full setup I tested costs about £930, although Tado lent me the hardware. Most of it can be installed yourself, but the boiler-wired part needs a heating engineer and typically adds around £80 in labour.

By default, the system uses push notifications to remind you to turn the heating off when you leave or when it detects an open window. To use the AI functions I tested, however, a subscription is required, costing £3.99 a month or £29.99 a year.

The paid tier is designed to improve efficiency through automation and extra features. The monthly subscription can also be paused when the heating is not needed.

Tado says its AI is trained using anonymised data from more than 1 million homes, which it does not sell to third parties. The system learns how quickly each room heats and cools, then uses that information to manage temperature and energy use more precisely.

It also carries out automatic hydronic balancing to improve efficiency, removing the guesswork from radiator balancing. It can switch off a radiator when it detects an open window, shut the heating down when the last person with the app leaves home, and preheat the house when it predicts someone is returning.

The subscription also stores 13 months of heating logs instead of just the latest month, and unlocks Tado’s Energy IQ tool, which breaks down heating costs overall and by individual room so you can identify possible improvements anywhere in the house.

Tado is Europe’s largest maker of smart heating controls, but alternatives with similar per-radiator control are available from Hive, Drayton and Honeywell. Google’s Nest, which sold smart-learning thermostats, withdrew from the European market.

AI versus manual

I programmed heating times for each room, enabled all the AI features and tracked monthly meter readings through the colder months.

In the first week, Energy IQ showed that one radiator was wasting a lot of heat warming an uninsulated porch. Since no one spends much time there before opening the front door, I changed the schedule quickly so it was mostly off.

After that, the system mostly ran on its own. At first, giving up control was difficult. Compared with my tightly managed manual routine, the boiler seemed to run more often, the rooms warmed faster and the house felt noticeably more comfortable. It was pleasant enough that I assumed energy use must have gone up. But to avoid being tempted to keep tweaking things and skewing the comparison, I avoided checking the figures as I went along.

When the totals were finally tallied, the result was a surprise: gas use had fallen by 18%.

Over the same 213-day period from November to May, with similar outside temperatures and after accounting for hot water use, the AI-controlled system used 11,735.7 kWh. That was 2,581 kWh less than the manual setup, even though the home was more comfortable. At the current price cap, that works out at a saving of £205.70.

Beyond efficiency

After two years of testing, my conclusion is that AI can reduce heating bills. In my case, usage dropped by 18%, and the savings could be even greater for households that are less careful to begin with or whose daily routines are less predictable.

There are also benefits beyond lower costs. Room-by-room control made it easy to keep the living room warm in the evening without overheating the bedroom overnight. I could warm a bathroom before an evening shower, heat only a child’s room at night, or raise the temperature in the porch to dry coats and boots in winter.

The temperature sensors and smart thermostats also measure humidity, making it easier to know when to run a dehumidifier or ventilate rooms to improve air quality.

All of that can be done while making the house significantly more comfortable and reducing the stress of managing heating and family disagreements over the thermostat.