Making Buildings Work Smarter – How Enviable Developed an Idea into a Working Service with Support from FAIR
Enviable set out to demonstrate that the underutilised real-time data generated by commercial buildings can be analysed with artificial intelligence and transformed into concrete actions that improve performance, reduce costs, and ultimately enable partially autonomous building management. With the support of FAIR, the company was able to validate its idea at an early stage and lay the foundation for a commercial product.
Text and photo by Eemeli Sarka, 25.5.2026
Over the course of his more than 30-year career in real estate, Enviable’s founder Mark Sorsa-Leslie repeatedly encountered a problem that no one seemed to be addressing: although vast amounts of operational building data were being generated, it was not being fully utilised.
Modern office buildings can produce terabytes of operational data every day. This data originates from HVAC systems, lighting, security infrastructure, and various environmental and occupancy sensors.
The real challenges, and opportunities, lie in interpreting that data, identifying meaningful anomalies, and translating insights into practical actions.
“There is enormous potential for optimisation, but we need a way to use the data effectively. And when no one else seemed to be tackling the issue, I decided to do it myself”, Sorsa-Leslie explains.
Can AI Understand What Is Happening Inside a Building?
The company’s original idea was to explore whether large language models could process real-time building data and identify operational issues. In collaboration with FAIR’s experts, Enviable launched a project designed to answer a fundamental question: does this approach work in practice?
“It was genuinely a Proof of Concept. We wanted to understand the limits of the technology and ensure that the original idea held up”, Sorsa-Leslie says.
The project demonstrated that the model works. The AI was able to detect anomalies in building data streams and interpret operational patterns in a meaningful way. Equally important, FAIR’s experts helped define what is technically feasible and what is not.
“The collaboration was fast and efficient, and the outcome was high quality. We knew what to focus on and what to avoid. In a short project, that makes all the difference”.
Working with FAIR also brought credibility. In discussions with customers and investors, Sorsa-Leslie can refer to the technical validation carried out by experienced AI professionals. In a market where slide decks alone are no longer convincing, a functioning demonstration is essential.
“It’s difficult to bring people on board with just an idea. It’s much easier to show how it works”, he notes.
The tangible prototype also helped attract a co-founder with a background in financial trading and energy systems, who brought deep expertise in real-time data processing and system architecture.
AI That Learns Continuously
Commercial buildings rely on building management systems that control HVAC, lighting, environmental and occupancy sensors, and security infrastructure. However, feeding raw data directly into a large language model is neither scalable nor efficient.
“For AI to work with real-time data, preprocessing is essential. You can’t just keep feeding unlimited data into the model. The solution has to scale across multiple buildings”, Sorsa-Leslie explains.
Enviable’s solution consists of two key components. First, real-time building data is converted into structured telemetry that AI models can process efficiently. Second, the system communicates identified issues to different stakeholders according to their roles – engineers receive technical explanations, while CFOs receive a financial perspective on the same issue.
One of AI’s advantages in detecting problems is that it does not rely on rigid rule-based logic. Instead, it identifies deviations within defined parameters that traditional systems might overlook. Detecting these subtle anomalies can lead to significant benefits.
“The energy-saving potential in a building is in the range of 20–30 per cent, particularly because AI can detect subtle inefficiencies that rule-based systems miss. For example, a ventilation unit may operate slightly above its optimal level without triggering an alert, yet over time this significantly increases energy consumption. AI looks at deviations from historical norms, not just fixed thresholds”, Sorsa-Leslie explains.
A Scalable Solution
Enviable is now preparing for the official launch and scaling of its service. The company focuses primarily on office buildings and works in partnership with facility management companies, engineering firms, and real estate consultancies.
Because the solution is purely software-based and designed for rapid onboarding, geographic expansion can happen quickly. Any building equipped with a building management system that generates data is a potential target.
“AI-based diagnostics and guidance allow engineers to manage more buildings with the same team or deploy junior staff more effectively. This increases productivity and enables additional billable work”, Sorsa-Leslie states.
For building owners, the financial impact can be substantial. An annual operating cost reduction of €50,000 can increase a property’s capital value by up to €1 million, depending on yield assumptions.
Broader ambition goes beyond cost savings. Moving from recommendations to automated adjustments and partial autonomy can improve energy efficiency, reduce emissions, and strengthen the operational resilience of buildings.
In an industry facing rising energy prices, increasing sustainability requirements, and growing skills shortages, scalable AI solutions may soon become not a competitive advantage, but a necessity.
