Property, facilities & energy
Buildings and grids are becoming software-controlled assets. Margin follows whoever controls them.
The lines that matter Operating cost per square metre and per MWh, availability and capital spend
Property owners, facilities operators and energy companies now run assets that software can monitor and control: buildings that adjust their own energy use, batteries that trade with the grid, and maintenance scheduled on condition rather than by the calendar.
At the same time, AI's own electricity demand is reshaping capital plans across Asia-Pacific. Regulators in the EU, the US and Australia are tightening rules on AI in critical infrastructure, and cyber agencies have issued joint principles for AI in operational technology.
- Rising operating costs per square metre and per MWh
- Ageing assets and large capital plans
- Sustainability reporting and energy-efficiency obligations
- Operational-technology security as more assets connect
- Building and asset data locked in separate systems
- Operational technology that is hard to connect safely
- Contracts that leave flexibility value with third parties
- Limited in-house data and analytics capability
Where AI can move the numbers
Opportunities grouped by Agent Value Ledger category. They are possibilities to test, not promised results.
Revenue
Energy flexibility revenue, tenant services, generation availability and trading
Operating costs
Maintenance triage, energy optimisation, field dispatch and crew planning
Capital
Condition-based capital planning, spare parts held on condition
Risk
Outage and fault detection, safety inspections, compliance reporting
The lines this sector is judged on
Property and facilities
- Net operating income
- Tenant service, lease administration, energy recharges
- Operating cost per square metre
- Maintenance triage, energy optimisation, cleaning schedules
- Asset value and capital spend
- Condition-based capital planning
- Compliance cost
- Safety inspections, sustainability reporting
Utilities and networks
- Operating cost
- Predictive maintenance, vegetation management, field dispatch
- Capital spend
- Targeting asset replacement and extending asset life
- Reliability incentives and penalties
- Faster outage detection and restoration
- Cost to serve
- Billing queries, hardship and collections
Generation and renewables
- Availability and yield
- Condition monitoring and maintenance timing
- Operating cost per MWh
- Fewer unplanned call-outs and better crew planning
- Trading margin
- Forecasting and bidding
- Imbalance cost
- Forecast accuracy against scheduled output
A maintenance agent across a wind portfolio
Assumptions: 500 MW of wind capacity at a 35% capacity factor (about 1.53 million MWh a year); realised price US$60 per MWh; unplanned maintenance budget US$6m; cost of capital 8%.
| Line | How the agent moves it | Measured by | Working | Per year |
|---|---|---|---|---|
| Revenue | Faults caught early, repairs timed for low-wind periods | Availability; generation | 0.5 points of availability × 1.53m MWh × US$60 | +US$0.46m |
| Operating cost | Fewer unplanned call-outs and better crew routing | Unplanned call-outs; cost per MWh | 15% × US$6m unplanned maintenance budget | +US$0.90m |
| Capital | Spare parts held on condition, not by rule of thumb | Spares inventory | US$2m released once, at 8% | +US$0.16m |
| Major failures | Component failures avoided | Failure rate against fleet history | Not counted until 12 months of failure history | Excluded |
| Cost of the agent | Model usage, sensor data work, engineer oversight, assurance | Total cost of ownership | Annual run cost | −US$0.50m |
| Net annual value | ≈ +US$1.0m, before any benefit from avoided failures | |||
Hypothetical illustration of the method. The figures are assumptions, not a client result or a forecast.
- Strict limits on what agents may control in physical and safety-critical systems
- Operational-technology security tested before agents are connected
- Human-reserved decisions for safety and emergency operations
Enterprise AI Transformation
Start with a diagnostic of maintenance and energy cost, then oversee a pilot with measured results on the ledger.
About this serviceEnergy & Sustainability in the Age of Agency
AI is now both the biggest new source of electricity demand and one of the best tools for managing it. Firms that treat AI agents as operational infrastructure, with matching controls, will run grids, buildings and portfolios at lower cost and risk.
Read the brief →Handing building and grid decisions to agents
Situation. A property owner's AI controls could trade the portfolio's energy flexibility with the grid, earning new revenue but handing real-time decisions to an agent.
Our approach. Clear limits on what the agent decides, operational-technology security tested against international principles, and a commercial model that keeps the value with the owner.
What good looks like. New revenue, tested controls, and a contract that keeps the value where it belongs.
Scenarios and worked ledgers are hypothetical illustrations. They do not describe specific clients.
Start with an executive conversation
A confidential discussion of your priorities, and where AI could move the numbers in your business.