Structured content management in the AI era

Zoltán is co-founder at Lombiq Technologies and an open-source maintainer

Organisations have spent years investing in content management systems, yet much of their content remains difficult to reuse, maintain and interpret outside the context in which it was originally published.

AI is exposing the consequences. Information that works perfectly well on a website may be much less useful when retrieved by an AI assistant, combined with other sources or delivered through an API. The words are available, but their meaning and relationships are not always explicit.

As Zoltán Lehóczky put it during our recent members call:

“AI does not magically understand messy content. It rewards content that already has shape.”

Zoltán is co-founder and managing director of Lombiq, a software development company based in Hungary specialising in open source content management. Lombiq has spent more than a decade contributing to the Orchard ecosystem, including Orchard Core, an open source CMS built on Microsoft's .NET technology.

Zoltán is also a Microsoft MVP and has been closely involved in developing the technology. His work with structured content goes back well before the current interest in AI. As early as 2015, he was already working on semantically structured content fields for Orchard, helping make information more meaningful, consistent and reusable.

During our conversation, he used Orchard Core to demonstrate principles that apply across the CMS landscape.

Better formatting doesn't solve the content problem

Using a blog post about space exploration, Zoltán demonstrated the progression from unstructured text to structured text and finally to structured content.

Adding headings, paragraphs and semantic markup makes content easier to read and interpret. Yet even well-formatted content can remain difficult to maintain and reuse when important information is buried inside a single text field.

The more consequential step is separating information into meaningful, independently managed elements. This allows content to be maintained consistently, connected to other information and delivered in different contexts without repeatedly recreating it.

For organisations managing content across multiple websites, channels and teams, this raises a question that goes beyond choosing a CMS: Are we managing information as reusable knowledge, or primarily as finished pages?

Create once, maintain once, use everywhere

Drawing on his work developing content management technology at Lombiq, Zoltán demonstrated how structured content can make reuse considerably easier.

Rather than copying information across multiple pages, Zoltán demonstrated how shortcodes and content references can help maintain a single source of truth, with updates reflected wherever that information is used.

His slides summarised the opportunity neatly:

“One content item, many outputs (listing, search, shortcodes, APIs…).”

During the discussion, our member David Kortleven from Plate CMS raised a related question about linking products mentioned in a blog post to the corresponding product records in a CMS.

Zoltán explained how content pickers allow editors to reference existing content items rather than manually copying information. This keeps the relationship intact and helps ensure that changes to the original information are reflected wherever it is used.

The organisational implications go beyond saving editors time. In many organisations, the same information exists in several systems, maintained by different teams. When something changes, it is not always clear who is responsible for updating it or which version should be trusted.

Structured content cannot resolve every question of ownership, but it provides a better foundation for managing information consistently.

Structure is about meaning, not just technology

Zoltán also demonstrated how semantic HTML and JSON-LD can describe content and its relationships in ways that search engines and other systems can interpret.

These approaches are increasingly relevant as organisations consider how their content will be discovered and used by AI applications.

Yet the underlying principles are not new. Semantic markup, content modelling and structured data have been part of good content management practice for years. What has changed is the number of systems that may need to consume the same information, often without the original page design or editorial context.

Zoltán emphasised that structured content is not limited to websites. The same thinking applies to knowledge base articles, press releases and even Word documents.

Zoltán captured the deeper challenge in one sentence:

“It’s not just about understanding content. It’s to understand how you intended it.”

A machine may recognise the words in a document without understanding why particular information was included, how different elements relate or which details matter most. Making those intentions explicit is part of what good content structure can achieve.

The work begins before AI

There is a temptation to approach AI readiness primarily as a technology challenge. Yet much of the necessary work begins with the content itself.

How is information organised? Where is it maintained? Who is responsible for keeping it accurate? Can it be reused without losing its meaning?

These are familiar questions for anyone who has worked with content management over the years. AI has made them more urgent, but it has not made the underlying challenges disappear.

Perhaps the most valuable shift is to stop treating content as something created for a particular page or document, and start treating it as knowledge that needs to remain understandable and useful in different contexts.

The conversation continues

Zoltán has spent many years helping organisations build structured, maintainable content systems that are designed for reuse. He recently joined us at CMS Summit 26 in Frankfurt and is an active member of the CMS Experts group. His experience adds another perspective to the conversations we're having about the future of content management.

Our members calls are one way we bring people together to exchange experiences and explore challenges across the CMS landscape. The conversations often continue in our peer groups and at our conferences, where there's more time to compare approaches and learn from one another.

You can also: