KPMG Finds AI Can Cut Building Energy Waste Up to 30%
AI-driven SEM can slash building energy use by up to 30%, beating one-off retrofits on cost and speed. Optimize first, replace second, add renewables after demand is stable.
AI Systems Can Cut Building Energy Waste by Up to 30%
What managers need to know
KPMG's latest research points to a clear priority: deploy AI-driven Strategic Energy Management (SEM) to cut building energy use by up to 30%. Efficiency gains at this scale typically beat one-off retrofits on cost, speed and portfolio impact.
The takeaway for management teams: optimise first, replace second, add renewables third. Getting consumption under control creates immediate savings and sets a stronger baseline for later capital projects.
Why efficiency-first beats retrofit-first
Traditional retrofits (boilers, chillers, insulation) help, but they won't get portfolios to net zero fast enough on their own. AI-led SEM can deliver fast reductions using existing assets, across climates and building ages.
As KPMG notes, renewables have limited value if the building is wasting energy. Optimise demand, then buy cleaner supply.
Real-world signal
"AI is already helping buildings cut waste by 20-30% in our projects, no matter the climate or the age of the property," says Donatas Karčiauskas, CEO of Exergio. "But those savings only last if there's smart energy management behind them."
The message is operational: savings come from ongoing oversight, not a one-time install.
The three-tier roadmap
- Tier 1 - Optimise what you have: AI tunes HVAC, lighting and controls in real time based on occupancy, weather and usage.
- Tier 2 - Upgrade equipment: Replace inefficient boilers, chillers and pumps once the baseline is stable.
- Tier 3 - Add clean supply: Onsite renewables and long-term power contracts after consumption is minimised.
How SEM with AI works
SEM alone typically delivers 5-7% annual savings. Combine it with AI and you reach 20-30%. The operating cadence is simple and repeatable:
- Assessment: Audit systems, data quality and comfort constraints.
- Planning: Set targets, budgets, and site sequence. Define comfort and safety limits.
- Implementation: Connect IoT and BMS, enable automated setpoint control, create exception rules.
- Capability building: Train facilities teams; define roles, alerts and escalation paths.
- Monitoring: Track KPIs, verify savings, adjust rules, and report monthly.
Think of it as creating a culture of active energy management. SEM sets the rules; AI keeps systems running to them minute by minute-while people stay in control.
Governance and transparency
AI must be explainable at the equipment and zone level. Facility managers should see the inputs (sensors, occupancy, weather) and the actions taken (setpoint changes, schedules) with clear reasons.
Establish guardrails for comfort, safety and peak demand. Use change logs and approval workflows for accountability.
90-day action plan for portfolio leaders
- Weeks 1-2: Pick 3-5 representative sites. Pull 12 months of utility data and BMS trends. Define comfort and uptime constraints.
- Weeks 3-6: Connect AI to HVAC and lighting controls. Enable automated setpoint/scheduling with human override. Stand up weekly review.
- Weeks 7-12: Validate savings with weather/occupancy normalization. Lock in rules that work. Build the rollout plan and capital roadmap.
KPIs to track
- Energy intensity: kWh/m² (or kBtu/ft²) versus weather/occupancy-adjusted baseline.
- Peak demand: kW reductions and demand charges avoided.
- Comfort: Time-in-range (% hours within temperature and IAQ limits).
- Interventions: Automated vs. manual actions, and issue resolution time.
- Verified savings: M&V reports (monthly and quarterly).
Procurement and IT checklist
- Integration: Supports your BMS/IoT stack; open protocols (BACnet, Modbus); read/write control with role-based access.
- Security: Encrypted data in transit/at rest; SSO; audit logs; network segmentation.
- Transparency: Explainability at point-of-action; change logs; configurable guardrails.
- Scalability: Multi-site management, policy templates, and portfolio dashboards.
- M&V: Built-in normalization for weather and occupancy; exportable reports for finance.
Risks to plan for (and how to handle them)
- Comfort complaints: Lock comfort bounds; monitor time-in-range and revert fast on exceptions.
- IT/security friction: Involve IT early; use approved integrations and network policies.
- Change fatigue: Train site teams; start with low-risk schedules; celebrate quick wins to build buy-in.
- Savings erosion: Set quarterly tune-ups; track drift; keep SEM reviews on the calendar.
Where to go deeper
Upskill your team
If you're rolling out AI across facilities and operations, ensure managers and site leads can work with these tools. Explore practical training paths here: AI courses by job role.
