I started where most people never look: audit files, plant rooms and control panels. As an ISO consultant I learned how organisations document their work. As a site engineer I commissioned building management systems at GSK Pharma, Aga Khan Hospital and DHA Suffa University.
Then I ran them. At UBL's head office I looked after the BMS, HVAC, UPS and lifts of a bank headquarters. I designed solar PV systems at Pantera Energy, and at Utopia Industries I automated power factor correction and led maintenance teams.
On every site, the same gap showed up. The BMS watches the chillers and air handlers, but the hundreds of split ACs in offices and meeting rooms are invisible to it. They run in empty rooms all evening and nobody knows. Putting a sensor and controller on every unit costs more than most facilities will ever spend.
The BMS sees the chillers. Nobody sees the split ACs. The camera on the wall already does.
So when AI became practical, I didn't start with a model. I started with what was already installed. EnergyCam uses the camera on the wall to watch the room and each AC's own display, giving split units the monitoring a BMS would, for zero hardware cost.
When I joined Disrupt (Gaditek), reporting a broken air conditioner meant scrolling a spreadsheet and hoping somebody saw it. I gave myself ten days and zero budget, just free AI tools, and built a digital twin of our building. Now you don't scroll; you click the broken thing. Noor, who walks every floor every day and knows the building better than anyone, sat down and helped build it.
Disrupt made a film about it. Engineer first, software second: every system on this page runs inside a working three-campus facility.