Making Content Ready for Intelligence

Unstructured content already holds the knowledge organisations rely on. The shift is learning how to reveal its structure so intelligent systems can use it.

Making Content Ready for Intelligence
© Rob Hoeijmakers


A short exchange on LinkedIn about llms.txt led me to meet Pieter Versloot from Plate. I expected a technical discussion about AI crawling behaviour. Instead, we ended up talking about the foundations of organisational memory.

How symbolic structure can emerge from the unstructured content organisations already produce every day.

That is the quiet breakthrough. Content doesn’t need to be forced into formal structure when it is created. The structure can be discovered. And once it is discovered, it can be maintained.

Meanwhile, publishing still serves humans. We write pages. We measure reach. That hasn’t changed. What has changed is who reads first. Language models digest our content before most people do. They internalise what we publish, form assumptions, and reuse that knowledge elsewhere.

Content behaves like knowledge. We still treat it like decoration.

When Bots Become Readers: Publishing in the Age of AI Crawlers
Listening to Matthew Prince on Azim Azhar’s podcast made me reflect on who actually reads my blog. People (like you), machines, or both.

Fragmentation as the default

As organisations scale, content spreads and slowly drifts apart.
Small copies accumulate. Facts diverge. Metadata becomes optional.

Editors continue. Machines continue.
Trust becomes fragile.

This is not a CMS problem. It is a knowledge integrity problem.

A conversation that connected the dots

Pieter approached this from the inside: a decade of building websites and CMS for large organisations. He saw publishing succeed at the surface while coherence failed underneath.

Plate’s response is not to replace systems. It is to reveal, repair and sustain what organisations already know, but cannot easily see.

The layer that sees

At the centre of Plate Delta is CAS, the Content Analysis Service.
It reads content across systems, notices patterns and detects gaps.

Repeated product names. Inconsistent facts. Duplicated assets. Tone shift across teams. Metadata that is present in one place and missing in another.

Machines do the auditing humans cannot scale.

“With Plate Delta, structure doesn’t come first, it follows the work. CAS reads how content is actually used and guides from real patterns, not preset rules. Structure stops feeling forced and starts doing its job.”
— Pieter Versloot

Structure emerges from reality rather than being mandated.

The layer that repairs and maintains

The Content Engineering Layer (CEL) then reinforces what CAS discovers.
It creates a continuous loop:

• detect divergence
• align to a shared model
• prevent new drift

Editors remain expressive. Systems remain flexible.
Coherence becomes infrastructural rather than personal.

“The bigger an organization gets, the more content rules drift into people’s heads or scattered tools. That doesn’t scale. You end up with systems that lock everyone down or with shadow IT that fuels chaos. Only a central, intelligent content base breaks that cycle.”
— Pieter Versloot

CEL does not centralise content.
It centralises alignment.

CEL acts as a smart layer on top of your existing systems, restoring control and clarity to your content management.

Where symbolic structure takes shape

Unstructured writing is how humans think.
From it, CAS and CEL surface:

• entities
• attributes
• relationships
• contexts and constraints

What begins as text becomes symbolic knowledge.
A knowledge graph forms organically out of use.

Meaning becomes computable.
AI can draw from a reliable foundation.

Rules and Guesses: The Twin Engines Behind Smarter AI
AI isn’t just smart, it’s a duet. One part rigid and logical, the other fluid and generative. Together, they form the neuro-symbolic systems behind real insight.

What this enables

A continuously maintained knowledge base improves daily work:

• faster editing, less hunting for inconsistencies
• reliable product information everywhere
• multilingual content aligned over time
• AI assistants grounded in verified truth

Structure becomes governance.
Consistency becomes trust.

Plate’s product set

Three components, one intelligence loop:

  1. Content Health Audit
    See what is broken
  2. CEL: Content Engineering Layer
    Repair and maintain alignment
  3. Plate CMS (Delta)
    Publish without reintroducing chaos

Machines reveal. Machines align.
Humans stay in control.

Looking ahead

Content becomes knowledge when it stays aligned over time.

And that memory becomes the internal logic on which every service relies.
If you recognise that your content is working against you as much as for you, there might be a better foundation available.

I’m not affiliated. Just impressed. And curious where this thinking leads.

Multi Site CMS & Digital Experience Platform | Plate
At Plate, we’ve been pushing the boundaries of content management for 10 years. Our mission is to give content and website managers full autonomy over their websites and content. Over 1,400 customers rely on Plate’s CMS products to navigate the fast-paced online world. Plate Multisite is the all-in-one solution for managing multiple websites, while Plate Delta is the headless CMS designed to structure and manage content seamlessly across various channels, all with minimal development needs and full content reusability.

What llms.txt can do for your website
When bots become interpreters of your brand, context matters. llms.txt helps guide what AI systems understand and repeat about your website.
The right CMS for your online content
Choose the perfect CMS: Learn what a content management system (CMS) does and find the right one for you.