AI-driven HVAC optimisation at Exchange Quay
Continuous fault detection across a multi-building office estate — measurable improvement without capital replacement.
Exchange Quay is a premier multi-building commercial office estate in Salford, Greater Manchester. With complex HVAC infrastructure and a diverse tenant mix, the estate faced the universal challenge of maintaining peak operational efficiency in a live, high-demand environment. A strategic partnership was formed between the asset owner TILL, managing agent CBRE, and CSR Sustain.
Client: TILL (asset owner) · CBRE (managing agent) · Location: Salford
- Control conflicts identified where heating and cooling were fighting each other
- “Ghost running” found — AHUs, pumps and fans operating during unoccupied hours
- Sub-optimal setpoints, control deadbands and plant sequencing corrected
- Invisible degradation spotted before it was severe enough to trigger BMS alarms
How the programme ran
Non-intrusive integration
The CIM PEAK AI-powered Fault Detection & Diagnostics platform was connected to the existing BMS with zero additional hardware and no tenant disruption.
Real-time AI analysis
Automated monitoring of thousands of data points to detect faults and abnormal operation as it occurred.
Expert oversight
CIBSE-certified Building Performance Engineers reviewed and validated every AI finding, ensuring recommendations aligned with best-practice engineering standards.
What made this difficult
Operational drift
Plant running outside occupancy schedules, with control settings drifting through seasonal change and manual overrides.
Conflict and inefficiency
Plant running outside occupancy schedules, with control settings drifting through seasonal change and manual overrides.
Reactive maintenance
Plant running outside occupancy schedules, with control settings drifting through seasonal change and manual overrides.
The data gap
Plant running outside occupancy schedules, with control settings drifting through seasonal change and manual overrides.
The outcome
Reduced HVAC energy consumption, enhanced plant longevity through lower runtime and mechanical stress, improved occupant comfort with fewer complaints, and a transition to proactive, data-led maintenance where performance drift is corrected in real time — all achieved without capital replacement.
Lessons
- The most sustainable building is the one already built — when it is optimised correctly
- Pairing AI diagnostics with certified engineering returns an estate to its original design intent
- Independence between the controls layer and the performance verification layer builds owner confidence
