This week we're featuring a blog from Giorgia Farella, an energy engineer, entrepreneur, and founder of Joynergy and Broken Pot, an ESCo active in energy efficiency and sustainable transition.
Giorgia has developed cross-sector experience across technology, energy, and innovation, with a specific focus on energy efficiency, decarbonization, artificial intelligence applied to energy, and behavioural models and joined us for the FIN Connect 2026 in Verona.
What do we ask of a building automation system today? For a long time, the answer was control: regulating systems, managing set points, supervising operations, and collecting data. Today, that answer is no longer enough.
Buildings are no longer merely places to be made efficient. They are assets under energy, economic, regulatory, ESG, and operational pressure. And they have become a concrete part of organizations' decarbonization strategies.
The future of BAS is not just automation. It is carbon-aware orchestration.
The question is no longer simply how do we make this building consume less? But how do we reduce emissions, prove it credibly, maintain comfort and operations, and preserve asset value?
For years, the market focused mainly on energy savings: reducing kWh, lowering costs, improving energy ratings, optimizing a few set points, and acting on the most energy-intensive systems. Then attention shifted to energy performance: not only consuming less, but understanding how the building consumes, with which profiles, under which conditions, and against which benchmarks.
Today, we are in a further phase: carbon performance. And where companies are not yet thinking about it, banks and customers often are.
A kWh is no longer just a kWh: it has a cost, but also a carbon intensity, a time of consumption, a source, an impact on peaks, a relationship with the grid, and contractual, regulatory, and reputational weight.
Two identical energy consumptions, at different times or in different countries, can have very different carbon implications. Energy efficiency therefore remains essential, but it is no longer sufficient to describe the whole problem.
To speak seriously about building decarbonization, we must include electrification, load management, flexibility, integration with renewables and storage, and measurement and verification of results. We must also consider the organization's ability to absorb new operational logics.
Decarbonizing a building does not simply mean installing a heat pump, adding photovoltaics, buying green energy, or introducing an AI-based solution. All these tools can be useful, but alone they do not make a strategy.
Building decarbonization is, first and foremost, a system of decisions: reducing demand, optimizing operations, electrifying loads where it makes sense, shifting flexible consumption when possible, integrating renewables and storage, and, throughout the process, measuring, verifying, and reporting. Where desired, residual emissions can be offset through voluntary carbon credits.
These steps influence one another. Installing renewables without first reducing demand can lead to oversizing. Optimizing only for cost, while ignoring carbon, can generate suboptimal decisions. Optimizing only for carbon, while ignoring comfort and operations, can create resistance. Implementing technology without governance can quickly bring the organization back to old operating habits.
In real projects, many buildings do not fail because they lack technology. They fail because technology is not governed over time. A building can be well designed and still operate poorly; it can have efficient systems and consume too much; it can have advanced dashboards and no resulting decisions.
The carbon gap is often an operational gap: the result of operational micro-decisions that are not coordinated, not measured, and not governed.
For this reason, the BAS of the future cannot merely control the building. It must become an operating layer capable of connecting system performance with carbon performance.
Field devices, HVAC, meters, and sensors remain at the base. Above them sits the BAS, with automation, supervision, and data integration. Then come analytics and AI: forecasting, predictive control, fault detection. But above that, a carbon-aware decision layer is needed: one capable of looking not only at kWh, but also at kg of CO2 equivalent, carbon intensity, demand peaks, comfort, actual occupancy, and costs.
The new optimization question is no longer “how do we minimize consumption?” but “how do we optimize energy, carbon, comfort, and operational constraints together?” This can mean pre-cooling when the carbon intensity of energy is lower, reducing peaks without compromising comfort, detecting HVAC anomalies before they become persistent waste, or preventing inefficient overrides.
This is where artificial intelligence can play an important role, provided it remains grounded in the reality of the building. AI is not magic. It creates value only if it survives the real building. And the real building is much less orderly than presentations: incomplete data, uncalibrated sensors, inconsistent naming conventions, systems that do not respond as expected, comfort constraints, maintenance habits, BMS limitations, and human overrides.
Algorithms, patents, applied research, and AI architectures matter. But field robustness matters even more: without data quality there is no AI; without control authority there is no optimization; without governance there are no results.
For this reason, AI should not be presented as the goal, but as a tool for making better carbon decisions: forecasting loads, weather, occupancy, and carbon intensity; optimizing HVAC setpoints, schedules, peaks, and operating strategies; detecting faults, drift, simultaneous heating and cooling; and, above all, explaining why a decision is being made. In HVAC, explainability is not a luxury: if an operator does not understand the system, sooner or later they will disable it or work around it.
When discussing decarbonization, the debate tends to focus on better systems, better controls, better algorithms, and better platforms. All correct, but incomplete.
Technological efficiency and AI create potential; behavioral efficiency converts that potential into results. This is not behavior in the simplistic sense of “turning off the light when you leave.” It is about how people interact with systems, how much they trust data, how operational decisions are made, which routines become established, and who owns the result.
A building is not only a system of equipment. It is a system of habits, constraints, and decisions. If people do not trust the system, they work around it. If they do not understand it, they ignore it. If the correct behavior is too complicated, they bypass it. From this perspective, behavioral efficiency is not energy moralism: it is the design of the interaction between people, systems, and decisions.
For all of this to become real decarbonization, MRV is also needed: measurement, reporting, and verification. A dashboard shows what happened; MRV proves whether an intervention worked, how much it worked, why it worked, and whether the result is credible. Without a baseline, emission factors, operational context, weather normalization, occupancy, and reference to business activity, we risk confusing correlation with results.
Finally, governance is needed. Someone must see the data, interpret it, decide, act, verify, and own the result. The facility manager governs operations, the energy manager interprets performance, the sustainability manager reports, finance evaluates investments and risks, operations define real constraints, and occupants influence the outcome every day. BAS can become the shared operating language among these worlds, but only if it is also designed as a decision-making infrastructure.
The future of BAS can be read as an evolution in five steps: control, measure, optimize, orchestrate, decarbonize.
Orchestrating means connecting energy, carbon, comfort, costs, maintenance, people, operational constraints, data, and accountability. It means making the best decisions not isolated episodes, but part of daily operations.
The future of BAS is not simply more technology, more AI, or more automation. It is carbon-aware, because it must also measure and optimize emissions impact. It is human-aware, because it must work inside organizations made of people, habits, and constraints. It is governance-aware, because without responsibility and verification, results do not last.
Decarbonization is not only a technical challenge. It is a measurement challenge, a behavioral challenge, and a governance challenge. And, above all, it is a challenge of orchestration.