AI in ISO 50001: How to automate energy reviews, EnPIs and baselines in practice

In brief

ISO 50001 requires energy reviews, EnPIs and baselines – but in large building portfolios, meeting those requirements manually is impossible. Here is how AI automates the core of energy management.

Contents

The strength of ISO 50001 is also its challenge: the standard requires continuous energy reviews, well-defined energy performance indicators (EnPIs) and documented energy baselines. In a portfolio of hundreds of buildings and thousands of meters, these requirements cannot be met with spreadsheets and manual data extracts.

This is where AI changes the game. When energy data, standard requirements and AI platforms are connected, the core of energy management can be automated – from data collection and anomaly detection to EnPI monitoring. This article shows how it is done in practice, based on our experience from large building portfolios and the EUDP project Next-gen Energy Management.

Why ISO 50001 and AI belong together

ISO 50001 is built around the Plan-Do-Check-Act cycle: planning with energy reviews and baselines, operation with targets and actions, checking with EnPIs and measurement, and acting through management review and improvement.

All four phases depend on data – and in large portfolios, the volume of data exceeds what people can handle manually. A municipality or company with several hundred buildings typically has thousands of metering points delivering hourly data.

AI platforms can analyse these data streams continuously: identifying abnormal consumption patterns, calculating normalised EnPIs and pointing to the buildings where action creates the most value. This turns ISO 50001 requirements from an administrative burden into a living management system.

Automated energy reviews in practice

The energy review is the foundation of the standard: where is energy used, what drives consumption, and where are the opportunities? Traditionally, this has required manual site visits, spreadsheets and static analyses – often updated only once a year.

With AI, the energy review can run continuously. The platform collects data from meters, BMS systems and building data, and automatically identifies anomalies: buildings heated outside opening hours, systems running at weekends, or consumption rising without explanation.

At the Danish Agriculture & Food Council, we have established energy management across 110,000 sqm with the AI platform Ento as the analysis engine. The work has earned us Best Practice awards from Bureau Veritas – precisely because the energy review was translated into concrete action.

EnPIs and baselines: From manual spreadsheets to continuous monitoring

Energy performance indicators (EnPIs) and energy baselines are the standard's measurement tools. But they are only useful if they are updated continuously and normalised for factors such as weather, opening hours and activity.

In practice, we often see EnPIs calculated once a year for the management review – making them useless for day-to-day operations. AI changes this by calculating EnPIs continuously and comparing actual consumption against the expected baseline in near real time.

At DTU Campus Service, we helped make ISO 50001 practical in a complex university operation across the campuses in Lyngby, Ballerup and Risø – with energy mapping, significant energy uses (SEUs), targets and action plans across more than 500,000 sqm of heated floor area. And at Fokus Nordic, focused energy screenings at four large commercial properties translated data into action: documented savings of more than DKK 500,000 annually, implemented through operation and control without major investments.

The EUDP project: Next-gen Energy Management

In the EUDP project Next-gen Energy Management, we are developing the future of AI-based energy management together with the project partners. The project explores how AI can translate standard requirements into concrete actions in operations – not just reports, but decision support used every day.

It is precisely the connection between ISO 50001's structured requirements and AI's analytical power that makes the difference: the standard ensures that energy management is systematic and documentable, while AI ensures it is current and action-oriented.

How to get started

The transition to AI-based energy management does not require changing everything at once. Our experience points to three steps:

• Get the data foundation right: Make sure meter data is available, valid and collected in one place.

• Start with the largest buildings: Automated analysis creates the most value where consumption is highest.

• Connect the analyses to your ISO 50001 process: Use AI output as direct input to energy reviews, EnPI monitoring and management reviews.

If your organisation holds an ISO 50001 certification – or is considering one – we can help assess how AI can strengthen your energy management in practice.