Butlr — Anonymous Thermal Occupancy Sensing

Product · HVAC & Energy

Product · InnDex 78 · Evidence provided · High specification risk

Thermal AI sensors that count occupants and optimise HVAC without capturing identifiable images.

Butlr's Heatic sensors use on-device thermal imaging and machine learning to detect and count building occupants in real time, enabling space utilisation insights and HVAC demand-response without storing or transmitting identifying imagery. Deployed across 30,000+ sensors in 100M+ sq ft globally across offices, labs, retail and senior living, the product addresses both operational efficiency and privacy-by-design constraints in occupancy-driven building management.

Butlr's Heatic sensors answer the occupancy question — how many people, where, right now — with on-device thermal AI that never captures an identifiable image, which dissolves most of the privacy and works-council friction that kills camera-based schemes. The deployment base is substantial for the category: 30,000+ sensors across 100M+ sq ft in offices, labs, retail and senior living, with one provided evidence item on record. The operational value arrives only when the count drives something: tied into the BMS for demand-based HVAC it attacks over-conditioning directly; standalone, it is a space-utilisation dashboard. Honest limits: thermal sensing degrades under high ambient temperature variance and direct sun, the on-device model is a black box whose counting accuracy and drift you cannot audit, and the energy-savings story leans on vendor claims rather than independent peer review. Pilot in a representative zone, validate counts against ground truth, and confirm the BMS integration path before scaling.

Strengths

• Privacy-compliant occupancy data without image capture, reducing privacy risk and regulatory friction
• Real-time occupancy-driven HVAC modulation, targeting energy waste from over-conditioned or under-used spaces
• On-device processing eliminates reliance on cloud inference, reducing latency and data egress
• Rapid deployment scale (100M+ sq ft) with named enterprise adoption signals market traction and interoperability credibility

Considerations

• Thermal imaging performance degrades in high ambient temperature variance or direct sunlight; not suitable for all building orientations or climates without performance trade-off (moderate)
• Requires integration into existing BMS/HVAC control layer; standalone sensor data has limited value without automated response infrastructure (moderate)
• On-device AI model is a black box to operators; limited transparency into counting accuracy, false-positive rates, or model drift over time (moderate)

Risks

• Thermal imaging accuracy degradation in high-humidity or extreme temperature environments (moderate)
• Reliance on on-device ML requires continuous calibration and updates across heterogeneous deployments (moderate)
• Limited independent peer-reviewed validation of energy savings claims (moderate)
• Sensor cost and ROI payback period not disclosed; adoption may be price-sensitive in cost-constrained segments (moderate)
• Thermal signature misinterpretation (e.g., heat-generating equipment vs. occupants) could lead to HVAC control errors (low)

Performance

• Global deployment scale: 30,000+ sensors across 100M+ sq ft
• Occupant detection accuracy: real-time counting capability
• Privacy compliance: no identifiable imagery captured or transmitted
• HVAC optimization: demand-response enabled

Reality check

Deployment scale (30k+ units, 100M+ sq ft, 22 countries) and customer names (CBRE, Lendlease, major tech firms) are plausible and consistent with Series B funding trajectory ($38M Series B, $68M total). Ricoh alliance signals enterprise traction. However: no independent third-party validation of claimed accuracy rates, anonymisation robustness, or cost-per-unit economics found. Privacy claims ('no identifiable image') lack published technical audit or certification. No published case studies with measured ROI or HVAC savings quantified. Fast Company award is credible but not a technical proof. Thermal imaging in high-ambient-heat environments (e.g. data centres, kitchens) may degrade; not addressed in available materials.

#occupancy_sensing #thermal_imaging #hvac_optimisation #privacy_preserving #ai_on_device #space_utilisation

Source