How can AI improve a building’s energy efficiency? The Smart Impulse x Foobot complementarity
In a commercial building (offices, healthcare facilities, industrial sites), excess consumption most often stems from an accumulation of small inefficiencies: restart schedules that are too broad, drifting setpoints, oversized ventilation at certain times, temporary settings that are never readjusted…
In this context, two levers make the difference when you want measurable savings while maintaining comfort:
- Understanding precisely where energy is going, by end use, using reliable, actionable data.
- Acting directly on building operations, especially the HVAC system (heating, ventilation, air conditioning), which is often a key driver of a building’s energy performance.
That is exactly where the Smart Impulse x Foobot complementarity comes into its own: an AI-based metering approach that makes consumption readable “by end use”, and an AI-based control layer that continuously optimises HVAC through the BMS.
Summary
- Measuring “by end use”: the foundation of an energy efficiency plan
- Using AI to control HVAC: Foobot as a “virtual specialist” connected to the BMS
- Why Smart Impulse strengthens Foobot: reliability, precision, proof of savings
- Client case: measure, control, maintain comfort
- ROI and scaling up: AI that turns intent into results
Measuring “by end use”: the foundation of an energy efficiency plan
The most useful question is not “how much does the site consume?”, but rather:
- What share comes from HVAC, lighting, IT, auxiliaries…?
- When do these end uses occur (working hours, night-time, weekends, off-peak periods)?
- Where are the operational drifts and anomalies that develop “silently”?
Smart Impulse addresses this need with a pragmatic approach: obtaining end-use analysis from a smart meter (Smart X) and AI-based disaggregation algorithms (NILM). The result is a clear, reliable view of electrical end uses, which becomes a solid basis for prioritising actions and evidencing savings.
Using AI to control HVAC: Foobot as a “virtual specialist” connected to the BMS
Foobot, a service marketed by EnergyWise, positions itself as an HVAC autopilot: an AI that makes fine-tuned adjustments to correct operational inefficiencies, while maintaining comfort.
In practical terms, which parameters does Foobot act on?
The AI controls a few key variables in the HVAC system, typically:
• heating and cooling curves (water temperature setpoints),
• air handling unit control,
• equipment start-up and shutdown schedules,
• or other parameters affecting consumption.
Foobot does not replace the BMS. It connects to it as an additional layer—interoperable and reversible (can be activated/deactivated).
Two key operational effects
- Continuous optimisation: the AI continuously detects and corrects small inefficiencies which, when combined, generate significant savings.
- Anti-drift: it automatically removes configuration drift (e.g. an “exceptional” setting left in place for weeks) that becomes costly over time.
Even in a well-managed building, operations naturally create deviations: successive human interventions, one-off adjustments, occupancy changes… and ultimately a gradual drift in energy consumption.
A very telling example: an exceptionally early start (a very cold Monday) can remain enabled all winter, even though it was only useful once or twice. Entrusting these parameters to AI helps avoid such energy drifts.
Why Smart Impulse strengthens Foobot: reliability, precision, proof of savings
Smart Impulse brings a very tangible advantage to the Foobot approach:
- Data reliability and availability
In many optimisation projects, the weak point is energy consumption data: missing, inconsistent, difficult to compare over time. When a site is equipped with Smart Impulse, electrical consumption is monitored, broken down by end use, and easy to exploit.
- Distinguishing HVAC from other end uses
Foobot targets savings on the HVAC system. Being able to isolate this end use (rather than estimate it) via an API with Smart Impulse makes results far more accurate—building by building, season by season—in two steps:
-
- Learning phase: consumption data from Smart Impulse helps establish a reliable baseline using a reference year for Foobot AI training.
- Results reporting: end-use reporting makes it possible to track savings, avoid misunderstandings about “what was really saved”, and clearly visualise performance over time.
Client case: measure, control, maintain comfort
A flagship example of this “measurement + control” approach concerns office buildings of around 13,000 m², where Smart Impulse and Foobot have been deployed.
Deployment steps: the “measure + control” logic
- Implementation of the metering plan with Smart Impulse (quick installation with no outage): clear end-use consumption reading enabling a precise separation of HVAC loads.
- Connection of Foobot to the BMS: an interoperable, reversible overlay.
- Learning phase: understanding building operation and thermal inertia using end-use data connected via API between the two solutions, then establishing the baseline.
- Continuous AI control: regular adjustments to key parameters, continuously identifying inefficiencies.
- Reporting: consolidation of savings and end-use tracking over time.
Observed results
27% savings were achieved on HVAC system energy consumption compared with the 2022 reference year—i.e. 170 MWh and around €60k savings “to date”—without degrading thermal comfort.
ROI and scaling up: AI that turns intent into results
In a context shaped by the Net Zero goal and stronger requirements around energy performance, many organisations are looking for solutions… but above all they decide based on proven ROI.
That is precisely where AI changes the game, not as a “gadget”, but as a driver of operational efficiency: it detects drift, continuously adjusts, and secures performance over time.
Fast ROI… because AI acts on the right levers
- Foobot: AI-driven HVAC optimisation targets an ROI of around 12 months, sometimes less depending on the building’s configuration and use.
- Smart Impulse: ROI is generally rapid, with an average 15% reduction in electrical consumption in under a year, because the tool immediately makes consumption readable by end use (HVAC, lighting, IT…), enabling action prioritisation, avoiding unnecessary investments, and turning assumptions into data-driven decisions.
That is also what makes the Smart Impulse x Foobot combination so operational:
- you measure better,
- you control more precisely,
- you prove faster,
- and you maintain performance over time.
Article written by Sabine Dorgan (Head of Marketing, Smart Impulse). Thanks to Inouk Bourgon (CEO of Energywise) his participation and contributions.
To go futher:
New Multisite Reports: a new feature for portfolio-level energy performance management
Energy Metering Plan: 6 steps to Smarter Energy Management