Can AI really save you money on your heating bill? I spent two years testing to find out

Can smart thermostats and AI reduce your heating bills? Every manufacturer claims the answer is yes, quoting average savings in the 25%, or £180 a year range. But all of these promises are heavily caveated and, with energy costs changing regularly, it can be very difficult to tell what is having an impact.

To find out if technology can really help, I spent the last two years comparing my miserly manual control and basic heating schedule to an AI-governed system. The results surprised me.

I used my 1930s solid-brick house with a recent, insulated loft conversion as the test bed.

Warming up

In the winter of 2024 my heating system consisted of a 10-year-old Worcester Bosch combination gas boiler set to a medium heat, a Hive thermostat in a central hallway with its smart features disabled and 13 radiators equipped with standard thermostatic radiator valves (TRVs).

To maximise energy savings, I set a basic schedule to turn the heating on for a couple of hours in the morning and evening, turned off the TRVs in rooms we did not use and manually adjusted those in other rooms when we did not need them to be heated. I also manually turned the heating off when leaving the house.

It was laborious and not all that comfortable at times. I frequently found myself being glared at by other members of the family who were shivering under blankets in the living room while watching The Traitors. Or being called a traitor after denying yet another request to turn up the heating. But it kept bills low throughout the winter and money in my pocket.

After controlling for hot water usage by using recordings from when the heating was not used during the summer months of 2025 to create a daily average, my manual control consumed 14,316.7 kilowatt hours of gas across 213 days from November to May. This was about 10 to 15% lower than in previous years when I had not been quite so controlling and miserly.

The switch

In the summer of 2025 I replaced the Hive thermostat with a full Tado X system connected to the existing combi boiler. This system consisted of 13 smart radiator thermostats to replace the standard TRVs, a wireless thermostat in each of the two largest rooms, and a programmer wired to the boiler.

Instead of a central thermostat and a schedule with just one temperature for the whole house, heating is managed room by room. Each smart TRV can control the temperature of an individual room by commanding the boiler to turn on and then adjusting the hot water flow into the radiator. The TRVs in rooms that don’t need heating turn the radiators off, which maximises the efficiency of the boiler as it has to burn less gas to maintain the same water temperature.

If your boiler supports the OpenTherm protocol – check your boiler’s documentation or contact the manufacturer to find out – it can also tell the boiler to turn down its flow temperature to reduce energy consumption. My boiler does not.

Tado’s smartphone app controls the lot and allows you to quickly mirror schedules across multiple rooms or get as fine-grained for each room as you like.

The system is not cheap. It costs about £930 for the full setup I tested – I was lent the equipment by Tado. It can be easily DIY installed apart from the part wired into your boiler, which costs about £80 in labour by a heating engineer.

By default, it uses push notifications to remind you to turn the heating off when you leave or if it detects an open window in a room. But for the AI-powered smarts I tested, a subscription is required, costing £3.99 a month or £29.99 a year.

This promises to further improve the system’s efficiency through automation and numerous additional features. The monthly subscription can be paused if you are not using the heating.

The AI is built by Tado using anonymised data collected from more than 1m homes, which it does not sell to third parties. It learns how quickly a room heats up and cools down and uses that information to more precisely control the temperature and energy consumption.

It also performs automatic hydronic balancing to increase efficiency, taking the guesswork out of balancing your radiators. It automatically turns off the radiators in a room when it detects an open window, turns off the heating when the last person with the Tado app on their phone leaves home and preheats your home for when it predicts you will step back in the door.

The subscription records 13 months of logs of your heating activity, rather than just the last month, and enables Tado’s energy IQ system, which shows you the overall cost of your heating down to the cost for individual rooms, helping to spot any potential for improvements in any part of your house.

Tado is the biggest manufacturer of smart heating controls in Europe, but there are alternatives with similar per-radiator control, including Hive, Drayton and Honeywell. Google’s Nest, which offered smart-learning thermostats, pulled out of Europe in 2025.

AI v manual

I set a schedule for when the heating would run in each room. I turned on all the AI features and recorded monthly meter readings throughout the winter of 2025 and into 2026.

In the first week, the energy IQ feature highlighted that one of my radiators consumed a lot of energy heating an uninsulated porch. Since you are barely in the porch for any time before closing the front door, it seemed a waste, so I quickly adjusted the schedule to turn it off most of the time.

After that the system quietly ran in the background without interference from me. Letting go was hard initially. Compared with under my miserly manual control regime, it seemed like the boiler was on more often, rooms heated up faster and the house was generally warmer and more comfortable. It was much more pleasant so I was convinced that we must be using more energy compared with the previous winter. But to avoid the irresistible urge to adjust things, which would bias the test, I took the “head in the sand” approach and did not run the numbers as I went.

When the final numbers were in, to my surprise we had used 18% less gas.

Across the same 213 days from November-May with relatively similar outside temperatures, and after adjusting for hot water usage, the AI-controlled heating used 11,735.7 kWh of energy, which was 2,581 kWh less than my manual control despite us being more comfortable. At the current energy price cap, ending on 31 December, that equates to a saving of £205.70.

Beyond efficiency

After two years of testing, I concluded that AI can save you money on your heating bill. We reduced our usage by 18%, but you could save much more if you are less frugal in the first place or lead a less predictable home schedule.

It also has multiple side-benefits beyond efficiency. The ability to control rooms individually made it very easy to have a warm living room in the evening without overheating the bedroom for the night. I could heat up a bathroom before an evening shower, increase the temperature of just a child’s room at night or heat the porch to dry out wet coats and boots in the depth of winter.

Each of the temperature sensors and smart thermostats also measures humidity, which made it easy to see when the dehumidifier needed to be put on, or rooms needed to be ventilated to improve the air quality.

And it can do all this while significantly increasing your comfort level, taking the stress out of managing your heating and your relationship with other family members.