Most business cases for smart building technology stop at energy savings. The bigger bills often come from things nobody tracks: faults that run for weeks before anyone notices, equipment that slowly drifts out of its commissioned state, and compliance and security risks that build up quietly, ruining smart building ROI.

In episode 48 of Sh*t You Wish Your Building Did, recorded live, we looked at how the cost of not knowing changes the way we should calculate smart building ROI. We were joined by Steve Watson of FPC Global, who has spent 35 years in the building industry split between facilities management and engineering and construction projects, and co-host Rob Murchison of Intelligent Buildings LLC.

Key Takeaways

Most buildings do not run the way they were designed. Without monitoring, they drift away from their commissioned state and become more expensive to run every year.
Fault detection and diagnostics (FDD) software has cut diagnosis time from about a day to about a minute. However, finding a fault creates no value until someone fixes it.
Data quality needs an owner and a budget for the whole life of the building, not just at handover.
Outdated building management system (BMS) firmware is a large, often unpriced cybersecurity risk.
AI is currently producing better dashboards. The real return comes when it helps close the loop from insight to action.
Improving building performance often requires zero capital investment. It starts with knowing what you have.

The Hidden Costs Missing From Smart Building ROI

Steve opened the podcast with a common example. During an audit of a 25-floor building, his team reviewed the time schedules for the heating, cooling, and air-conditioning plant in the BMS. The schedules were “all over the place,” with excess run time on each floor ranging from 1 to 17 hours, as schedules had simply been set to run 24/7.

The energy waste was significant, but the more telling finding came next. Once the problem was identified, the facilities management (FM) team did not have enough understanding to fix it immediately. Nobody knew whether the conditions that caused the schedule changes still existed. That disconnect between how a building is used and how it is configured is a real, recurring cost, and it rarely shows up in a smart building ROI calculation.

Rob asked whether most buildings run the way they were designed. Steve agreed that they did not. Design intent tends to get “put on the shelf” once a building is handed over.

Finding Problems vs. Fixing Them

Rob raised the central question of the episode: are we investing to find problems, or to solve them? If a fault is found but never fixed, the investment returns nothing.

Steve said the industry now spends a lot of time finding problems, and that connecting building data to a smart building platform has made detection dramatically faster. Good FDD tools have taken diagnosis from a day to a minute, a difference of roughly three orders of magnitude. Typical findings include:

Simultaneous heating and cooling
Valve hunting, where a valve constantly overshoots its target and shortens equipment life
Schedule drift and incorrect manufacturer settings on floor-level air conditioning units
Energy demand peaks that nobody has visibility into

Steve’s big-picture view is that buildings are constantly sliding toward disorder. Unless maintenance effort keeps pace with the rate of deterioration, running costs rise. That drift is visible in the data, and the most useful thing an FDD platform can tell an owner is how far the building has moved from its commissioned state.

Smart Building ROI as a Risk Management Problem

Rob framed the cost of not knowing as a two-sided risk. On one side is the risk of not knowing what is happening in your building. On the other is the risk that someone else gets access to your building data, for example learning who was on which floor at what time.

Steve gave a concrete example: a scan of BMS controllers in a recent building found firmware that had not been updated for around 10 years. He said this is not an unusual story, and it creates a large attack surface. It is hard to put a price on this kind of risk until it happens to you, which is exactly why it gets left out of the business case. Rob argued that facilities operational technology (OT) teams and IT teams must work together to manage it.

Where AI Delivers Return Today

Steve identified two ways AI is already returning smart building ROI:

Faster time to diagnosis and repair. Previously, an FM team asked to cut energy use by 10% had to pull data from siloed systems, such as lighting controls and the BMS, in different formats and protocols. Connected platforms now make that fast and simple.
Capturing expert knowledge. Much of the reasoning behind a repair is never recorded. Large language models can document diagnosis and reasoning, and that knowledge can be attached to specific equipment or repair categories through knowledge graphs.

Looking further ahead, Steve wants to run building simulations at scale, for example testing what happens to total cost if chilled water temperature drops by one degree, run across a vast number of scenarios.

Do We Still Need Dashboards?

Rob asked a deliberately provocative question: are dashboards on their way out? Steve went further. If an AI agent is running a building, does it need BMS or lighting control graphics pages at all? Probably not.

James noted that most buildings are nowhere near that stage. AI is mostly being used to produce better data for better dashboards. Until AI closes the loop by making and acting on decisions, its full potential is untapped. Steve added that simply improving trust in the data is a major step, because validating data consumes a great deal of time and effort.

Rob linked this to the OODA loop (observe, orient, decide, act) developed by US Air Force Colonel John Boyd. AI is excellent at the first two steps, which is the “find” side. The industry spends most of its effort there and not enough on deciding and acting, and without those steps nothing gets fixed.

The Costs Nobody Puts in the Business Case

Asked which costs are most often left out, Steve pointed to ongoing data quality. Organizations invest in designing and building the infrastructure, get it working at commissioning, and then stop investing in the data. He sees this becoming a new role within FM teams or an outsourced service, with proper funding assigned. He doesn’t think the cost is large.

He also expects AI to make updating building information models (BIM) easy within one to two years. He believes the as-built model, the computer-aided facilities management (CAFM) system and handover documentation should all align, and much of the governance to support that already exists.

Advice for Building Owners

Asked what an owner should do next week to improve smart building ROI:

Steve Watson: Understand your building. Identify the gaps in knowledge, skills, and complexity that your team can’t keep up with, then look at the tools now available to close them. Not knowing how your building runs is costing you “an absolute bomb.”

Rob Murchison: Getting a building to run better likely requires zero capital investment, so don’t start by buying the shiny new thing. Good decisions start with knowing what you have, how it’s performing and what’s changing. He also argued that governance, a concept long established in IT, needs to move into the OT world so all building technology can be governed as one.