How can AI help us organize and create policies for building operations?
This LinkedIn post starts this important conversation
The LinkedIn post you shared hits on a major challenge facing modern facilities: technology is moving faster than our rules for managing it. As buildings introduce automated systems and AI agents, the gap between what should happen (policy) and what actually happens (operations) expands.
To organize and create effective policies for your building operations—especially while bridging the gap between human oversight and autonomous systems—you can use a step-by-step approach modeled around the Trust Matrix concepts discussed by the panel.
1. Map Out Your Trust Matrix
Before writing policies, understand who (or what) is doing the work and who is holding the rules. Break your operational needs into four distinct zones:
- Human-to-Human (The Traditional Zone): Policies written by management and executed manually by your staff (e.g., standard maintenance checklists, manual temperature overrides).
- Human-to-Agent (The Automated Frontier): Policies written by humans but executed by smart Building Automation Systems (BAS) or AI (e.g., automated overnight setbacks, dynamic air-quality flushing).
- Agent-to-Agent (The Future Zone): Systems communicating with other systems (e.g., your energy management AI buying power or shedding load based on automated utility pricing).
2. Standardize the “What Should Happen” (The Policy Layer)
A policy is only useful if it can be verified. Instead of vague guidelines like “Keep the building comfortable and efficient,” your policies need to be machine-readable, explicit, and grounded in industry standards.
1.Adopt Proven Baselines:Step 1.
Don’t reinvent the wheel. Use existing industry definitions like ASHRAE Guideline 36 (high-performance sequences of operation for HVAC systems) as your legal and technical foundation.
2.Define Hard Boundaries:Step 2.
Clearly write down your constraints. What are the absolute minimum/maximum acceptable temperatures? What are the indoor air quality (IAQ) metrics that must trigger fresh air intake? These become the programmatic rules an autonomous agent can check.
3.Codify Accountability and Hand-offs:Step 3.
Define exactly when a human must override an autonomous system, and how that hand-off is logged. If an AI agent changes a fan speed or valve position, it must leave a clear digital paper trail.
3. Close the Gap with Continuous Auditing (The Operations Layer)
The panel mentioned a harsh reality: many buildings have guidelines on paper, but almost none actually comply day-to-day. To prevent your new policies from becoming shelf-ware:
- Implement Automated Fault Detection and Diagnostics (FDD): Use software that constantly checks real-world sensor data against your policy boundaries. If a valve is stuck open or an AI sequence is running inefficiently, the FDD system highlights the divergence immediately.
- Establish a “Source of Truth”: Keep all sequences of operations, setpoints, and building rules in a centralized, updated digital repository (like a digital twin or a unified configuration file) that both your engineering team and your software vendors reference.
The Takeaway: Building a trustworthy, modern facility doesn’t mean jumping straight into full autonomy. It means creating robust human governance first, translating those rules into clear parameters for your building systems, and building an audit loop to make sure your operations always match your intentions.
