The problem
The air-conditioning equipment in the company’s buildings was producing a steady stream of readings: temperatures, airflow, equipment status. None of it was collected anywhere it could be analysed, so there was no way to see where energy was being wasted.
What we built
Building equipment speaks industrial protocols, not web ones. We connected to it through KEPServerEX, which reads the equipment over OPC UA, the standard language of building and factory machines, and passed the live readings to AWS IoT SiteWise.
In SiteWise, each air handler and zone became a digital twin: a model of the real equipment that keeps its readings organised and in context. From there the data lands in an S3 data lake, AWS Glue cleans and reshapes it, and Amazon Athena lets it be queried like an ordinary database.
The client-facing result was a QuickSight dashboard showing how the system behaves over time. The design took account of how these systems actually run: air handlers feeding variable airflow to each zone, and demand that rises and falls with occupancy and the weather.
What comes next
The proposed next step is a model, trained on this data, that adjusts the air-conditioning to save energy while keeping people comfortable. That part has not been built.
If your project is similar, mention it in your brief and we can go into more detail in writing.