Commercial buildings waste roughly 30 percent of the energy they use, creating losses that drain property budgets and make climate targets harder to reach. The problem gets worse in residential towers, where aging HVAC systems run on outdated schedules, leaks go unnoticed for weeks, and maintenance workers can’t access sensor data trapped in equipment that won’t share information with other systems.
Property managers in cities face a tough call. Replacing outdated building-management gear costs millions, but ignoring the problem burns cash and puts buildings at risk as cities write tougher emissions rules. New York’s Local Law 97 started fining buildings that exceed carbon limits in 2024, and Boston, Washington, and Los Angeles have passed similar laws. The real question isn’t whether to upgrade, it’s how to connect systems built in different decades without shutting down operations.
Ashok Kumar Kalnayam, a technical program manager, spent several years building a solution that skips the expensive replacement route and uses software to make old equipment talk to new platforms. He worked with a property-management company running hundreds of residential units along the East Coast, leading the rollout of what the industry calls a Building Intelligence Operating System, a cloud setup that collects information from chillers, meters, door locks, and repair tickets and puts it all in one place for analysis.
Making old systems work with new software
The hard part wasn’t installing sensors. It was getting equipment from different eras to communicate. Building-management systems installed in the 1980s and 1990s use protocols like BACnet or Modbus, which were built to link a few devices in the same room. Modern internet-connected devices expect different formats and send data over newer channels. His design used edge devices at each property to translate the old signals into formats the cloud could read, then moved that information through secure connections to AWS storage.
“We couldn’t shut anything down, so we ran both systems side by side for three months while the old vendor contracts ended,” Ashok Kumar Kalnayam said. “Running them together let us fix bugs like one chiller that sent temperature readings in tenths of degrees when our software expected whole numbers, before anyone lost heat or hot water.”
The platform started pulling meter readings every fifteen minutes instead of once a month through manual checks. Maintenance crews spotted problems that had gone unnoticed: a boiler working at 90 percent capacity on a warm day, a lobby air system turning on and off sixteen times an hour instead of four.
The money saved came faster than expected. His team tracked a $400,000 drop in yearly energy costs across all properties, mostly from stopping equipment that ran when buildings sat empty and catching steam-trap failures before they turned into expensive emergencies. Service requests that used to take three days to close now finish in under two days because dispatchers can see which worker is closest and whether the needed part is already on site.
Writing the rules to keep it running
Technical setup is half the work. He also wrote procedures that spell out who manages each piece of data, how often equipment gets checked, and what problems trigger an alert versus just a note in the log. Property companies often end up with systems built by three or four different vendors over the years, each with its own login and reporting schedule. Bringing that mess together meant writing contracts that promise uptime levels and explain what happens when a gateway stops working.
“The maintenance team didn’t trust it at first; they’d watched test projects work fine in demos but break down under real use,” Ashok Kumar Kalnayam said. “We convinced them by putting up a live status page anyone could check and making sure the new help process was faster than the old one, not slower.”
That doubt makes sense. One survey found 68 percent of smart-building projects fail because facility workers don’t get trained on new software or because the system needs constant support from the vendor. Ashok built training into the schedule, short weekly sessions where technicians practiced running reports and closing tickets in the new system, and set up a help desk that answers questions within two hours.
The setup also handles compliance reporting. As cities add stricter emissions rules, the central database can produce records showing energy use by the hour, refrigerant refills, and how long equipment ran without anyone hunting through files from six contractors. That matters as more places adopt building performance standards through this decade.
What other industries are noticing
His work sits where two shifts in building management meet. Cloud storage prices have dropped enough that keeping years of sensor data makes financial sense; AWS and Azure now price IoT services in ways that work for mid-size building portfolios. At the same time, fewer people are entering skilled trades. Bureau of Labor Statistics projections show HVAC technician numbers will fall 5 percent through 2031, so systems that catch problems before they need emergency repairs are gaining users.
Data centers are paying attention. They use about 200 terawatt-hours of electricity each year in the United States and are testing similar software to improve cooling and power systems. Factories with older control equipment face the same connection problems, and some are adapting methods from real estate to manufacturing floors.
He has heard from hospital groups interested in adding patient-room controls, temperature, humidity, and air circulation to the same platform. Healthcare buildings must meet Joint Commission rules, and being able to print timestamped compliance reports quickly appeals to administrators who currently track everything on paper.
The next step involves using machine learning for predictive maintenance. With two years of data now stored in the cloud, His team is testing algorithms that recognize vibration patterns before pump bearings fail or predict which air filters will clog early based on pollen levels. The models are still experimental, but early results suggest they could cut reactive maintenance costs by another 10 percent.
None of this required new technology. Ashok’s contribution was recognizing that the toughest problems in smart buildings are about people and process, aligning vendor contracts, training workers who’ve seen failed upgrades before, writing procedures that survive when staff leave, and that technology only works when it makes someone’s job simpler instead of adding steps. Buildings waste energy not because sensors don’t exist, but because the people managing properties are stretched thin using tools that can’t share information. Fixing that needs software, but it also needs someone willing to sit through budget meetings where return-on-investment numbers matter more than technical diagrams.
