Engineering and asset data

The path to AI readiness.

AI is usable in operations when the same question gives the same answer every time, and every answer can be traced to its source.

That comes from the data, not from the AI.

What it takes

A map beneath, and an address on every building.

Industrial asset
In-house analysis and AIOpen questions, repeatable answers
3.Asset mapStructured knowledge: usable by AI
2.Common languageDS/EN IEC 81346: one designation system
1.Source dataDocuments and databases, unconnected
The foundation an AI needs
Analogy: getting around
GPS navigationAny destination, reliable directions
3.The mapConnected roads: usable by GPS
2.AddressesCountry, city, street, number
1.Roads and buildingsOut there, but unlisted
Foundation for navigation

From source data to AI-ready asset knowledge. IEC 81346 provides the common language. Beside it, the same three layers in a setting everyone knows.

Three layers · One structure the AI can use directlySame reason GPS navigation works
1

Source data

The material already held in documents and databases: diagrams, manuals, maintenance records, spare parts lists, requirements. Complete in each system, connected in none.

2

Common language

IEC 81346 gives every technical object one identity, wherever it appears. Not a new naming scheme, but an address system that the existing names can be resolved to.

3

Asset map

Documents, properties, requirements and maintenance data connected into one structure the AI can use directly. It is the same reason GPS navigation works.

Why indexing the documents is not enough

The material is there. Nothing connects it.

Indexing the documents gives the first layer and only the first.

The same pump can appear under four different names, and nothing in the text says they are the same object.

An AI reading that is working from photographs of buildings. It can describe each one, but it cannot give directions between them, because no photograph contains a street.

And when it fails to connect them, nothing shows. The answer is less complete, but it still reads well.

What it makes possible

Ask in plain language. Get the same answer every time.

With the three layers in place, an AI can work directly on the asset. Staff ask in a chat window, and get answers that hold.

  • What is installed where, and what will a change affect?
  • What is missing or undocumented?
  • Which work can share one shutdown, and what carries the greatest consequence?
  • And, in principle, any other question the asset data can answer.

Every answer points back to a component, a document and a revision, and all of it runs on the organisation’s own systems.

Equipment with no document

The map knows what belongs in the area it covers, so it can list what is installed but never described.

Documents with no equipment

Files that refer to nothing that exists any more, or that never existed under that name.

Requirements with nothing recorded

Demands that were set, and against which no test, inspection or document was ever filed.

A system built on documents alone cannot do that. It has no way of knowing whether something is missing or simply was not found.

Where to start

One area of the asset. The data you already have.

The whole asset does not have to be described before the first questions can be answered.

Build from what exists

The foundation is built from the documents and databases already in place, one area of the asset at a time. Nothing has to be re-documented first.

See it before you decide

The principle can be shown, not only described. A preliminary model of one area is enough for a short demonstration of how this works in practice.

Keep what holds its value

An AI can be added at any time, and whichever one is added will be replaced by a better one. The map of the asset is the part that keeps its value.

An AI can only be trusted if the data beneath it is structured and traceable.

The OntoTeq way of looking at data

Your data knows more than you think.

The asset map is not new data. It is the structure already implied by the documents, tags and records you have, made explicit and connected. OntoTeq reveals that structure, and Know-Y is where it becomes navigable and, from there, AI-ready.

Is your asset data ready for AI?

Start with one area and the data you already have. We’ll show you what the map already knows, and what it says is missing.