The Starting Point: High Costs, Low Visibility
Across the portfolio, the symptoms were familiar:
- Rising demand charges and a steady drumbeat of hot/cold complaints.
- Fragmented data: utility bills living in finance, BMS data trapped in vendor silos—no shared “source of truth” or clear why behind monthly peaks.
- Alert fatigue: engineers were drowning in alarms, leaving little time to fix root causes.
The team didn’t lack effort. They lacked leverage—specifically, a way to see the whole system and act on it continuously.
The Shift: From Set-and-Forget to Closed-Loop
We implemented a four-step playbook designed to turn raw data into measurable outcomes—automatically.
1) Unify the Data Layer
We integrated utility interval data, submeters, and BMS/IoT points into one normalized stream.
This single pane of glass did more than visualize—it made the data queryable, comparable, and auditable across sites.
2) Let Agents Do the Heavy Lifting
AI agents surfaced specific, profitable actions—think sequencing fixes, economizer logic, condenser ΔT issues, and rogue schedules. Engineers weren’t asked to “do more.” They were handed rank-ordered work tied to real loads and real dollars.
3) Automate the Wins
Where changes were deterministic and safe, we automated:
- Pre-cool on hot days to flatten late-afternoon peaks.
- Demand-shed during events with targeted, rules-bound strategies.
- Schedule & setback enforcement building-by-building to eliminate drift.
4) Close the Loop (Continuously)
Every change was verified in live data. If it saved, it stayed. If not, it rolled back. That feedback loop turned improvement into a 24/7 process, not a once-a-year project.
What Changed: Outcomes You Can Measure
Within the first season, the portfolio saw:
- 15–20%+ reduction in electric costs, with even bigger wins at sites dominated by demand charges.
- Peak kW shaved via targeted shed and smarter pre-cool logic.
- Fewer hot/cold calls as sequences stabilized and drift disappeared.
- Actionable M&V: every savings claim tied to tagged points and intervals, making finance and operations equally confident in the numbers.
This wasn’t “AI for AI’s sake.” It was process discipline—codified, automated,
Why It Worked
Two truths made the difference:
- AI didn’t “think” for the building.
It watched patterns and enforced better decisions, consistently, at machine speed. No heroics. No magic. - Control of digital infrastructure and data.
Because the owner owned the network, the data model, and access to building systems, we could normalize points, apply agents, and automate without waiting on vendor silos. Data rights and plumbing weren’t afterthoughts—they were the foundation.
What This Looks Like for Your Team
- Engineers spend less time chasing alarms and more time delivering durable fixes.
- Asset and property managers see a clear line from projects → savings → NOI.
- Finance gets defensible, interval-level M&V—no soft math.
- Residents/tenants experience steadier comfort with fewer edge-case failures.
If you’re still tuning HVAC by hand or hoping last year’s schedules hold, you’re paying an invisible tax. Closed-loop optimization turns that tax into a dividend.
A Simple Way to Start (30-Day Sprint)
- Centralize the last 12 months of utility intervals + current BMS/IoT points.
- Normalize & tag the top 200–400 points per building (economizers, supply/return temps, VAV status, schedules).
- Pilot agents at 2–3 representative sites (office, medical, multifamily)—focus on sequencing, schedules, and demand events.
- Automate only what’s deterministic; keep humans in the loop for exceptions.
- Verify savings weekly in live data; publish a one-page M&V recap to leadership.
Result: a credible, repeatable playbook—not a one-off project.
Credit & Context
Many of the insights here were sharpened in conversation with Mical Anselmo (NODA.ai) on Peak Property Performance — Ep. 12, hosted by Bill Douglas & Drew Hall. The through-line from that episode is clear: own your digital infrastructure, normalize your data, and let agents turn intent into outcomes.
Check our Peak Property Performance Linkedin for more updates on the podcast and book
Ready to reclaim your energy spend?
OpticWise helps owners unify data layers, deploy closed-loop optimization, and prove savings with point-level M&V.
If you’d like the pilot checklist or a sample M&V one-pager, say the word—we’ll share the template.
Let’s start with your PPP Audit. Contact us today.