Property owners can derive greater value from existing data by structuring information and clarifying the relationships between buildings, installations, and metering points. Graph-based information models can complement traditional databases, creating a better foundation for AI, more efficient operations, and well-informed decision-making.
We have plenty of data. But how much do we understand?
Imagine a typical office building.
It has building control systems, energy meters, temperature sensors, air handling units, property management systems and a range of other digital solutions.
They all collect information.
But if you want to understand why energy use suddenly increases, you often have to start searching.
Which air handling unit serves which spaces?
Which data points belong to that unit?
Have the operating hours changed?
Is there a connection between energy use and a change in the tenants’ activities?
The information is often already there. But it is not always connected in a way that makes these questions easy to answer.
We have become good at collecting data. But we have not always been equally good at describing how things fit together.
And those connections are essential when we want to move from traditional monitoring to more advanced analytics and AI.
Tables describe information. Graphs bring relationships into focus.
For a long time, relational databases have been the foundation of how we organise digital information.
We store buildings in one table, equipment in another and data points in a third.
Different identifiers allow us to link the information together.
It works. And it will continue to work.
But as reality becomes more complex, it also becomes more important to describe and explore the relationships between objects.
This is where graph-based information models offer an interesting alternative.
In a knowledge graph, we describe objects and how they relate to one another.
A building contains an air handling unit.
The unit serves a number of spaces.
A sensor measures a temperature within the unit.
An energy meter records its energy use.
Suddenly, we have more than a collection of objects with properties.
We have a connected description of how the building works.
The difference is not that tables lack relationships, but that graph models make those relationships an explicit and central part of the information structure.
This can simplify the work of exploring connections, combining information from multiple systems or developing new digital services.
An important distinction: graph database or knowledge graph?
The terms are sometimes used as if they mean the same thing.
But a graph database is a technology for storing and managing graph-structured information.
A knowledge graph describes objects, their meaning and their relationships.
The latter is particularly interesting for the property sector.
Because what matters is not which database we choose.
What matters is establishing a shared understanding of the objects within a building and how they fit together.
A knowledge graph can also be built alongside existing systems. Not all information needs to be moved to a new platform.
From isolated systems to a connected understanding of the building
Imagine that we want to analyse energy use in a large building.
With a traditional approach, we retrieve meter readings from the energy management system, operational information from the building control system and details of the building’s equipment from other systems.
Then we need to connect the information.
But what if we have already described the relationships between the building, its equipment, data points and spaces?
The same information can then be used by several different services.
An energy application can analyse consumption.
A maintenance management system can identify the equipment involved.
A building operations technician can trace the relationships to understand an anomaly.
And an AI assistant can be better equipped to retrieve the right information for an analysis.
We move from each application having to understand the building independently to multiple applications using a shared description of it.
That is a significant change.
Particularly when we want to scale digital services from a single building to an entire property portfolio.
AI changes what we need from our information structure
When we discuss AI, we often focus on which language model is best.
Which one can analyse the most information?
Which one can automate the most tasks?
But for a property owner, there is another question that may be more important:
How can AI analyse a building if it does not know which sensor belongs to which air handling unit?
An AI model can process large amounts of information and identify patterns.
But it does not automatically know how a particular building works.
It needs information about objects, their functions and their relationships.
Let us return to the energy example.
We want to ask:
Why did energy use increase in this building during September?
To investigate that question, AI needs access to more than a list of energy readings.
It needs to identify which equipment affects consumption, which data points belong to that equipment and what changes have occurred.
For example, an air handling unit’s operating hours may have changed, or activities in part of the building may have changed.
A shared information model makes it easier to retrieve relevant information and investigate possible connections.
Of course, a knowledge graph does not guarantee accurate analysis. The information must be quality-assured, and the conclusions must be verifiable.
But it can give AI something very valuable:
Context.
AI can be highly capable at analysing data.
But we still need to give it the foundations to understand what that information represents.
Information needs to be reusable
Another important consideration is that information should not be locked into a particular application or supplier.
Describing a building’s objects and relationships in a shared information model creates new opportunities.
The same structure can support energy optimisation, operational analysis, maintenance planning and future AI services.
We can develop new solutions without having to describe the entire building all over again.
This is where open information models and shared standards become interesting.
One example is Brick Schema, which describes building systems and equipment and the relationships between objects.
The technology itself is not the most important part.
The principle is what matters.
Knowledge of how a building works should be a shared resource — something that does not have to be recreated in every new system.
This gives property owners greater flexibility and a stronger foundation for developing their digital ecosystem over time.
From data to understanding — the real technological shift
We will continue to use tables.
Relational databases will remain an important part of our digital infrastructure for a long time to come.
But as our digital solutions become more advanced, we also need to develop the way we describe information.
We need to move from simply knowing which objects exist to understanding how they fit together.
From storing readings to understanding what they represent.
From separate sets of information to connected knowledge.
This is where knowledge graphs and modern information models can play an important role.
Not as a replacement for all existing technology, but as a way to make information easier to understand and use.
Because the digital buildings of the future are not just about collecting more information. They are about understanding the information we already have.
And perhaps that is where we need to start before investing in the next major AI solution.
Next: digital twins — but what are we actually building?
We are talking more and more about digital twins.
Advanced 3D models, visualisations and platforms that bring together information from different systems.
But what actually turns a digital representation of a building into a digital twin?
Is it an impressive 3D model?
Is it connected, real-time data?
Or is it something far more fundamental?
Can we really create a functioning digital twin without first understanding the relationships between the building’s different parts?
In the next article, we examine the concept of digital twins and discuss what they actually need to deliver value in property management.
At PropNodeX, we connect, structure, validate and manage property information. Our starting point is that information should remain usable and reusable over time, independently of individual systems and suppliers.