Facility Management: Reliability, Planning, and What AI Can and Can't Do

Facility management is fundamentally a reliability discipline: keep the building systems running, know what's about to fail before it does, and make defensible capital decisions with limited budget. Everything below reflects how I've actually run this work, from multi-site commercial portfolios to mission-critical federal facilities today.

Preventive Maintenance

A PM program is only as good as its follow-through. Building the schedule is the easy part; the hard part is knowing, in real time, which PM tasks are actually being completed versus which are quietly slipping. Most reliability failures I've seen trace back to a PM task that was scheduled but never verified.

Asset Lifecycle Planning

Every major mechanical asset -- chillers, boilers, air handlers -- has a real, knowable lifecycle. Planning replacement before failure, rather than reacting to it, is the difference between a budgeted capital project and an emergency purchase order at the worst possible time.

Work-Ticket Analysis

Work tickets are a genuine data source most facilities teams under-use. Patterns in repeat tickets on the same asset are usually the earliest real signal of a developing failure -- long before it shows up as an emergency.

Recommissioning

Recommissioning -- verifying that building systems are actually operating as designed, not just as installed -- is where I spend a significant part of my current work. It routinely surfaces drift between how a system was designed to run and how it's actually running today, and that gap is almost always where efficiency and reliability are being lost.

Capital Planning

Capital improvement planning is a coordination problem as much as a budgeting one: engineering recommendations, real asset condition, and available capital rarely arrive on the same timeline. Defensible capital planning depends on having real asset-condition data ready when the budget conversation happens, not scrambling to justify a request after the fact.

Operational Reliability

Reliability is the sum of the above done consistently, not a single initiative. It's built from PM follow-through, honest asset lifecycle tracking, and a real feedback loop from work tickets back into planning.

Vendor Coordination

Coordinating vendors, service teams, and facility-improvement projects across multiple sites is genuinely difficult to do from memory. The teams that do it well have a real system tracking who owns what, not a shared inbox.

Mechanical-System Oversight

Mechanical system oversight across a portfolio means knowing the condition and priority of every major asset at every site simultaneously -- something that becomes genuinely hard to hold in one person's head past a handful of sites.

Multi-Site Facilities

Multi-site operations multiply every one of the challenges above. What works informally at one site breaks down at five, because informal tracking doesn't scale with headcount or geography.

Government & Mission-Critical Environments

Mission-critical facility work -- my current focus, supporting operations at MacDill Air Force Base -- raises the stakes on all of the above. There is less tolerance for reactive maintenance and a higher bar for documented, defensible operational decisions.

Where AI Can Help Organize and Prioritize

The genuinely useful role for AI here is organizing and prioritizing information a facilities team already has -- surfacing overdue items, aging patterns in work tickets, and follow-ups that are falling through -- so a human can make the actual engineering call faster and with better information.

Where Human Engineering Judgment Remains Required

AI does not make engineering decisions, does not certify equipment, and does not replace a qualified engineer's assessment of a system's condition. Any tool -- including Metrix Command Center -- that claims otherwise should not be trusted with facility operations.