Manufacturers of smart home technology frequently claim that artificial intelligence can slash heating costs by roughly 25%. A two-year independent test conducted by a Guardian journalist suggests these figures are plausible, though the actual savings depend heavily on previous habits. The study found that an AI-governed system reduced gas consumption by 18% compared to a highly disciplined manual approach.
Manual Control vs. AI Automation
The test took place in a 1930s solid-brick house with an insulated loft conversion. During the winter of 2024, the homeowner relied on a basic schedule and manual adjustments to keep costs down. This involved turning radiators on and off based on room usage and frequently adjusting settings to maintain warmth. While this method kept bills low, it resulted in a cold living environment and friction within the household.
Data from this period showed a consumption of 14,316.7 kilowatt hours (kWh) of gas over 213 days. This figure was already 10 to 15% lower than previous years due to the strict manual control. The homeowner used summer months to establish a baseline for hot water usage, ensuring a fair comparison for the next phase of the experiment.
In the summer of 2025, the system was upgraded to a Tado X setup. This included smart radiator valves, wireless room sensors, and a boiler programmer. Unlike the previous central thermostat, this system managed heat room by room. It utilized anonymized data from over one million homes to learn how quickly specific rooms heated up and cooled down. The AI also performed automatic hydronic balancing to optimize boiler efficiency.
Measurable Results and Comfort Gains
The AI system required a subscription for its advanced features, costing £3.99 per month. It offered features such as open window detection, geofencing to turn off heating when the last person left, and preheating before arrival. The system also provided detailed energy logs, allowing the user to identify inefficient areas, such as an uninsulated porch that was wasting energy.
After running the AI system through the winter of 2025 and into 2026, the results were recorded. The AI-controlled heating used 11,735.7 kWh over the same 213-day period. This represented a saving of 2,581 kWh compared to the manual control phase. At the prevailing energy price cap, this translated to a saving of £205.70.
Despite using less energy, the household reported higher comfort levels. The ability to heat specific rooms independently allowed for a warm living room in the evening without overheating bedrooms. The system also tracked humidity, helping to improve air quality and reduce the need for dehumidifiers. The test concluded that AI heating savings are real, particularly for households with less predictable schedules or those willing to be less frugal initially.
The full Tado X setup tested cost approximately £930, with an additional £80 for professional installation of the boiler wiring. While the upfront cost is significant, the annual savings and improved comfort may offset the investment over time. Other manufacturers like Hive, Drayton, and Honeywell offer similar per-radiator control, though Google Nest exited the European market in 2025.
Source: The Guardian
