Over the past two years, virtually every conference has been talking about AI. Artificial intelligence will control buildings, optimise energy consumption, detect faults and predict failures.
It sounds very convincing. There is just one problem: I think we are completely unprepared for it.
And it is not about AI at all. The buildings themselves have proved to be the biggest problem.
I expected a completely different problem
In recent years, I have been developing my own HVAC control platform, creating a methodology for integrating building services and data exchange protocols, and testing these approaches in real commercial buildings.
To be honest, I thought artificial intelligence itself would be the greatest challenge. I was wrong.
We build buildings that cannot talk
A modern commercial building very rarely functions as a single system. More often, it consists of dozens of separate worlds: VRF from one manufacturer, ventilation from another, fire safety controls from a third and a BMS from a fourth.
Meters operate separately. Lighting separately. Access control separately. Energy monitoring separately.
Each has its own protocols, logic, database and interface. In practice, we end up not with a digital building, but with a collection of systems that barely interact with one another.
AI is powerless here
Many people today say that AI will control buildings. I think they are right. But there is one problem.
Artificial intelligence cannot analyse what it cannot see.
If temperature data is in one system, energy consumption in another, equipment status in a third and alarm logs in a fourth, AI does not have the full picture.
It receives isolated pieces of information instead of a model of the building.
And this is not just a Ukrainian problem
You might think this situation is unique to our buildings. Unfortunately, it is not.
Working with different systems, I became increasingly convinced that the problem is much broader. Huge numbers of buildings around the world have building services that work well individually but barely interact with one another.
That is why the most important investment today is not a new sensor. Nor even a new algorithm. Integration is becoming the most valuable investment.
A new design brief
I believe developers should already be asking designers a new question.
Not: ‘What equipment should we install?’
But: ‘Will this building be able to work with any HVAC Intelligence platform in ten years’ time?’
This is a completely different way of thinking. A building must be designed not only for people, but also for future digital control systems.
Key takeaway
I believe that in the coming years, competitive advantage will not belong to the building with the most expensive equipment. Nor even to the one that consumes the least energy.
The buildings that handle data best will win. Those where all building services can exchange information. Those where every parameter can be measured.
Those where artificial intelligence sees not an individual air conditioner or pump, but the whole building as a single organism.
The key question for the future is no longer ‘Do we need AI?’ It is much simpler: ‘Is our building ready for AI at all?’
Once the data sources and quality are understood, the next step is a limited, verifiable task. An example AI pilot for building services shows which signals are needed to analyse out-of-hours operation and how an engineer checks the recommendation.