From distributed data to coherent intelligence

Monday morning, 08.17 am. A leaking roof at a housing corporation. An outdated installation at a water, industrial or healthcare organization. Overdue maintenance of municipal real estate. The problem differs, but the question is the same: 'Repair, replace or immediately make it more sustainable?'

Let's zoom in on the housing corporation. That's where everyone takes action: the maintenance worker sounds the alarm. The contractor takes an inspection report from an old project folder. An experienced colleague refers to two previous leaks that have not been properly recorded anywhere. The real estate manager plans roof replacement in three years. The sustainability consultant wants to insulate the complex. And in the meantime, finance monitors the budget.

Everyone has a piece of the truth. But where is the complete story? The good news: this is present. But how do you bring all the pieces of information together when you have to make a crucial decision?

The leak has been fixed... the problem is not

Most companies opt for the practical solution: solve the problem, book costs, keep going. They focus on the things in front of them: a work order, an inspection report, an investment rule, an energy goal, a contract and a project schedule.

But is repairing actually still wise? Can you make the repair of the roof coincide with planned renovation and insulation? What implications does postponement have exactly, for example for living comfort, energy consumption or resident nuisance? And is this an incident or does this indicate a pattern?

A water company, industrial organisation or healthcare institution has a similar question about an installation: 'Are you going to repair or replace it again and combine replacement with a more energy-efficient alternative?' And a municipality must consider whether maintenance of a building, road or sewer system can be linked to sustainability or a planned area approach.

The assets differ, but the underlying consideration is: 'What should we do now, what can wait and what can you combine smartly?'

To answer these kinds of questions, you need a coherent picture of the situation, use, history, costs, laws and regulations, sustainability and timing. That is what many companies lack. They work with separate knowledge elements that are spread over systems, documents and the heads of experienced employees. Every single piece of information is correct, but remains on an island. That way you don't come to a well-considered decision.

What do you need? Context. And you create that by combining and managing the available knowledge.
 

Bringing all the separate knowledge elements together: how do you do that?

Information only becomes meaningful when you can always answer this question: 'What does this piece of data say about which object, at what level, at what time?'

The first step is not to unleash AI on this. Before you do this, it is smart to create a knowledge hierarchy in which you give each area, complex, building, system and component a recognizable place. This hierarchy must be in line with your organization.  For example, in the case of a housing corporation, an inspection is attached to a roof, a renovation plan is attached to a complex and a sustainability goal is attached to the entire portfolio. In other organizations, an inspection may be attached to a pump, a policy rule to an area and a performance profile to an installation.

You then link three types of sources to that structure:

• Structured data: work orders, inspections, contracts and budgets.
• Unstructured knowledge: policy documents, manuals, permits and project files, as well as the recorded experience of operators and mechanics.
• Current signals: alarms, measured values, energy consumption and the status of ongoing projects or permits.

With this approach, you don't create a new data silo. Your source data will remain where it is managed. But the context layer questions and indexes them, giving them more meaning.
 

What is the role of AI?

Where the context layer connects information, AI ensures that you can actually use that knowledge. And that goes beyond just answering questions. With agentic AI , multiple specialized AI agents can work together to retrieve information, analyze it, and translate it into concrete advice. One agent determines what information is needed, while other agents consult relevant sources, analyze data, and compare different scenarios.

Take the leakage at the housing corporation. AI can bring together and analyze maintenance history, planned renovations, costs, and sustainability goals. This will provide insight into whether repair is wise in the short term or whether replacement in combination with insulation will ultimately yield more.

But AI can also go a step further. Agents can prepare follow-up actions, support processes, and assess new signals. Think of identifying recurring malfunctions or preparing a maintenance proposal.

Data governance plays an important role in this. Because how do you ensure that AI uses the right information and that advice remains reliable and verifiable? Techniques such as retrieval-augmented generation (RAG) allow agents to access relevant sources. Good data governance ensures clear agreements about data quality, access rights, topicality and traceability. This provides insight into the sources, versions and assumptions on which an advice is based. Facts and AI-generated conclusions remain distinguishable from each other and the final decisions lie with humans.

The context layer gives information meaning. AI translates that meaning into insight, advice and action. Data governance provides the frameworks within which this is done responsibly.
 

Three roles, three questions, one source of information

What does this approach yield? That differs per role. To illustrate this, we highlight three of them! In doing so, we show which questions each role faces and how one source of information answers them from one reality.
 

Director or asset owner

How do you keep an outdated portfolio affordable while the challenges for quality, continuity, safety and sustainability are only growing?

A housing corporation is about the relationship between maintenance, quality of living, sustainability and room for investment. In the case of a water company, industrial company, healthcare institution or municipality, the performance and facilities for which that organisation is responsible.

With the context layer, the director or asset owner sees more than what is planned. There is insight into the interventions you can cleverly combine, the places where risks accumulate and the actual costs of postponement.
 

Asset and portfolio manager

Where should the next euro go, and why?

The context layer shows where poor condition coincides with high impact, which programs serve multiple purposes at the same time and where replacement actually comes too early. For example, a housing corporation makes it clear at which complexes you can tackle roof maintenance and insulation together. Other organizations may see that they can combine installation replacement with energy savings. Or that they can bundle maintenance within an area program.

In addition, it is possible to trace any advice back to the source data and the assumptions behind it. As a result, the asset and portfolio manager no longer has to make decisions based on gut feeling.

Operations and Maintenance Manager

How do we minimize rush jobs, avoid unnecessary repetition and move towards predictable performance?

The context layer immediately enriches every report or malfunction with the history, the network position, the contract agreements and similar assets elsewhere in the organization. If there is a recurring problem (and therefore a pattern), the operations and maintenance manager sees it much more easily.
 

Where do you start?

You might want to connect all your sources first and clean up each document. Our advice: don't start with this.

Rather, choose one decision that comes back again and again in your company. For example, are you going to repair, replace or combine maintenance with sustainability again? Whether it concerns a housing complex, an installation or municipal real estate: build the context around it that all parties involved need. When this foundation is in place, you immediately get value from it. And from there, you can further expand the context layer.

Curious about the best decision to start with? Feel free to contact us to discuss this.