The future of digital construction, autonomous buildings, and AI-driven building management systems offers a scalable solution. By combining digital twins, AI, and collaborative operational platforms, the industry can unlock unprecedented energy savings, improve compliance with new EU regulations, and enhance asset value.
Troy Harvey, CEO at PassiveLogic, in his keynote at PropTech Connect 2025, highlighted how there is a paradigm shift in how we design and operate the built environment. His company is, he says, disrupting the world of automation with the first fully automated platform for buildings, reinventing automation with real-time control decisions and in so doing reduce a buildings carbon footprint.
“Automation in buildings has long promised efficiency and sustainability, but its impact has remained limited. Today, fewer than 3% of large-scale commercial buildings use advanced automation systems, leaving the remaining 97%—particularly those under 10,000 square metres—largely untouched. In many U.S. and European markets, building automation still means basic thermostats, relays, or even manual radiator valves.
“This gap matters because buildings are the single largest contributor to global energy consumption and emissions, accounting for nearly twice the energy use of both transport and industry. Without intervention, the building sector will continue to drive climate change at an unsustainable rate. The good news is that up to 40% of building energy use could be saved immediately through better control systems, without the need for extensive retrofits.”
Why building automation hasn’t scaled
“Despite decades of innovation,” he said, “automation in the built environment has remained a niche solution.” Adding that several barriers can explain this namely:
- Fragmentation: Multiple siloed controllers and cloud platforms make integration difficult.
- High Costs: Legacy building automation systems are often too expensive for small-to-medium buildings.
- Complexity: Every building is unique, requiring bespoke programming and commissioning.
- Slow Data Cycles: Buildings only generate one full seasonal data cycle per year, limiting AI training.
“The result is that building controls remain stuck in “open loop” systems, failing to deliver the intelligence required for large-scale climate and cost impact,” he elaborated
The regulatory push: EPD compliance in the EU
In Europe, the Energy Performance of Buildings Directive (EPD) is forcing rapid change. By 2026, all EU member states must comply with new standards that require minimum control and analytics capabilities in buildings.
Key thresholds include:
- 2026: Compliance required for buildings above 290 kW.
- 2029: Threshold drops to 70 kW—covering buildings of around 1,000 sqm or more.
Penalties are steep. In Spain, fines can start at €60,000 for non-compliance, escalating by €1,000 per week. Authorities also reserve the right to install control systems at the owner’s expense.
This makes investment in digital building platforms not just a sustainability decision, but a financial imperative.
Digital Twins: the foundation of autonomous buildings
“To overcome the limitations of legacy systems, companies like PassiveLogic are pioneering the use of digital twins in construction and building management,” states Troy.
“And unlike traditional BIM models or cloud-based databases, digital twins integrate multiple capabilities namely:
- Visualisation (BIM-based): Representing the physical structure.
- Time-series data integration: Capturing live building performance.
- Physics-based simulation: Predicting outcomes and testing scenarios.
- Neural net inference: Enabling AI-driven control.
- Cross-domain normalisation: Standardising data for real estate, logistics, and industrial applications.
“Therefore by creating a virtual model that “knows” the physics of pumps, boilers, HVAC systems, and electrical equipment, digital twins enable AI to simulate decades of building performance in hours—bypassing the slow learning cycles of real-world data.”
AI and differentiable programming in building operations
Traditional programming follows a rigid input-to-output model. In contrast, differentiable programming—the backbone of modern AI—reverses this relationship, allowing systems to reason backwards from outcomes to inputs.
In practice, this means:
- Deductive inference: Predicting future building states from current conditions.
- Abductive inference: Determining the optimal control strategies to reach a desired future outcome.
- Inductive inference: Estimating unknown conditions based on limited sensor data.
Together, these capabilities enable AI-powered building management systems that don’t just monitor performance but actively optimise it in real time.
From point solutions to autonomous platforms
“Most current smart building solutions are fragmented—single-purpose IoT devices or isolated cloud services. The future lies in integrated platforms that bring all systems together under a unified digital twin,” he explained.
Stating that such platforms provide:
- Plug-and-play interoperability with manufacturer-provided digital twins.
- Automated commissioning by drawing the building and its systems rather than writing code.
- Continuous learning as the building evolves with occupancy, equipment ageing, or tenant changes.
- Custom AI agents for tasks such as predictive maintenance, CO₂ reduction, energy optimisation, and tenant engagement.
“This turns the building into an autonomous workforce, capable of managing itself while providing actionable insights to owners and operators,” he added.
Benefits of autonomous buildings
Investing in an AI-driven building automation platform stated Troy delivers multiple advantages:
- Energy Efficiency: Up to 40% savings without retrofits.
- Regulatory Compliance: Seamless alignment with EPD and other global standards.
- Asset Value Uplift: Transforming B- or C-grade buildings into A-grade assets, boosting valuations by up to 10%.
- Tenant Experience: More responsive and comfortable environments.
- Operational Savings: Reduced truck rolls, predictive maintenance, and optimised resource allocation.
In conclusion Troy said: “The construction and real estate sectors stand at the threshold of a profound transformation. While traditional automation has failed to scale, digital twins and AI-driven building platforms now offer a path to widespread adoption. With EPD compliance deadlines looming and the climate crisis intensifying, autonomous buildings represent not just an innovation but a necessity.
“By treating buildings as dynamic, intelligent systems, the industry can unlock massive energy savings, reduce emissions, and future-proof assets. In doing so, the built environment moves from being the world’s largest energy consumer to its single greatest opportunity for climate action.”
Read next: RoboBallet sees robots learn to work together – Build in Digital
Are you a building professional? Sign up for a FREE MEMBERSHIP to upload news stories, post job vacancies, and connect with colleagues on our secure social feed.
