A long time ago, I worked as a student mentor, helping students with subjects like math, history and Dutch. It was a great side job because it allowed me to revisit these topics at a time when I could actually enjoy them. One day, I was helping a student with Dutch and got asked the question: What is context?
Well, imagine walking into a supermarket dressed as a pirate in the middle of autumn. People might give you some strange looks. But if you dress up as a pirate during Carnival, people will smile at you and hand you a beer. During Carnival, the context is right.
Context makes all the difference. And the same applies to AI.
The next generation of AI in manufacturing won't be won by whoever has the most data, but by whoever gives AI the right context.
What happens when AI detects irregularities, but doesn't have the context to understand them? What if a technician halts a production line based on a false alert?
For example:
Without the right context, AI might identify a problem where there isn't one, or miss the actual cause of an issue.
For an AI agent, context is the complete set of instructions, information and environmental data available to the model at any given moment. It helps the agent reason, make decisions and take action.
This is the foundational layer that determines whether an AI agent acts as a capable collaborator or makes confident but incorrect decisions based on hallucinations.
This is also the core architectural principle behind Intelligence Center X (ICX): AI that meets your data where it lives, using knowledge graphs.
Instead of loading all information into a single data lake, a knowledge graph connects underlying data and systems.
ERP, WMS and MES systems all speak different languages. Manually mapping the data between them is time-consuming and error-prone. In one system, something might be called a "part", while another refers to it as a "component". This makes it challenging to compare and understand data across systems.
A knowledge graph provides the network of relationships between all these elements. Taxonomies and ontologies give this network structure and meaning. This allows AI to understand what things are and how they relate to each other. For example, it can recognize that a part and a component refer to the same thing.
When AI truly understands the relationships between orders, materials and machines, you can start unlocking real value:
The value lies not just in having access to more data, but in understanding the relationships behind it.
As a manufacturer, you don't have to make this transition overnight. You can take a phased approach, as many of our customers do.
Here's what those phases could look like:
At Squad Apps, we help manufacturers make this transition step by step. We combine human-centered design, expertise in knowledge graphs and ICX, and low-code platforms like Mendix to build solutions that your shop floor operators will actually use.
Understanding is what unlocks the real power of AI and agents. If you use AI that doesn't truly understand what it's doing, you might just be moving really quickly in the wrong direction.
That's why we believe in the power of Intelligence Center X.
Curious about how knowledge graphs can bridge the gaps between your plant's different data systems? Let's connect and share ideas!
