Predictive heating control is moving beyond single-building demonstrations in the Nordic property market, with Siemens and Swedish AI company Algeno deploying their combined system across 73 buildings for nine customers in Sweden and Norway.

The partnership connects Algeno’s predictive technology with Siemens Building X, using operational information, weather forecasts, and observed building behaviour to automate heating decisions. The companies are targeting existing property portfolios, where energy systems may already produce substantial data but replacing underlying building-management infrastructure can be expensive and disruptive.

Siemens says many Nordic buildings still rely on fixed heating schedules and manual adjustment, making it difficult to respond efficiently to changing weather, occupancy, and operating requirements. The companies’ approach is designed to sit above existing systems and use data already available from the building rather than requiring owners to replace their controls before introducing optimisation.

Although the announcement gives a useful measure of deployment scale, it does not publish aggregated figures for energy consumption, emissions, tenant comfort, or financial savings across the 73 buildings. That leaves an important part of the commercial case unresolved: whether a system that is easier to deploy also produces sufficiently consistent returns across different properties.

Existing buildings carry the harder problem

New commercial buildings can be designed around modern sensors, controls, and connectivity from the start, but much of Europe’s property estate predates those systems. Heating plants, building-management software, meters, and environmental controls have often been installed in layers over many years, leaving landlords with a mixture of equipment and data formats.

Replacing that infrastructure to support an AI project can make the economics unattractive before optimisation begins. Siemens and Algeno are therefore emphasising an open-platform architecture intended to let Algeno’s software integrate through Building X across several sites without a large bespoke engineering exercise at each property.

Heating depends on more than outdoor temperature. Different buildings retain heat differently, usage patterns change throughout the day, internal equipment contributes heat, and weather forecasts can allow a control system to act before conditions change rather than reacting afterwards.

Predictive systems try to learn those relationships and adjust heating earlier or more precisely than a fixed schedule. In practical terms, the model may determine when heating should begin, how aggressively a system needs to respond, or whether a building can coast through part of the day without compromising operating conditions.

For landlords, a software layer can potentially improve the performance of assets that will remain in service for decades. Property owners cannot wait for an entire portfolio to be rebuilt before reducing energy use, while heating costs remain exposed to weather, energy prices, and regulatory pressure around building performance.

Automation shifts responsibility into software

Allowing software to make heating decisions changes the operational model. A facilities team that previously altered schedules manually is increasingly supervising rules and outcomes produced by an automated system, so property managers need confidence that the technology responds sensibly when occupancy, equipment, or weather deviates from historical patterns.

Override mechanisms and performance monitoring remain important even when optimisation is highly automated. Excessive heating wastes energy, while underheating can produce tenant complaints, damp problems, or operational disruption. A fault elsewhere in the mechanical system can also make a software recommendation appear ineffective even when the model is behaving correctly.

The 73-building deployment should provide a broader base for judging those issues because a portfolio creates variation that a single pilot cannot reproduce. Offices, residential blocks, public buildings, and other properties differ substantially in thermal behaviour, maintenance standards, control equipment, and occupancy.

Scale also tests whether a platform remains manageable. A landlord may be comfortable with engineers closely monitoring one experimental building, yet the economics change when a small facilities team is expected to oversee dozens or hundreds of automated sites. Exception handling, remote configuration, and integration with existing maintenance processes become as important as the prediction model.

Techopia recently examined another attempt to turn building and energy data into maintenance signals, reflecting a wider shift in property technology away from collecting information for dashboards and towards systems that trigger operational action.

The climate case will ultimately depend on measured consumption rather than the presence of AI. Heating is a major component of building energy use, so effective optimisation can reduce waste, but gains vary with the condition of the building, the efficiency of its plant, existing control quality, and how poorly it was operated before the software arrived.

Siemens and Algeno have reached an important but incomplete stage: nine customers and 73 buildings establish that the technology can be deployed across a portfolio, while the absence of published aggregate savings leaves its broader economic and environmental performance to be demonstrated.