For years, Confluence was relatively easy to understand conceptually. You had a space, inside that space you had pages, and pages were essentially the unit of knowledge. The page tree gave information a place and, even when it became messy, the mental model remained fairly clear.
That model is starting to change.
Confluence now contains pages, Live Docs, whiteboards, databases and slides, alongside an increasingly important AI layer through Rovo. Information can be created in different formats depending on whether the goal is to document, collaborate, explore, structure or present it. The boundaries between those formats are becoming less rigid, and that changes the way knowledge itself is organized.
Slides are a particularly interesting addition to this picture. A presentation has traditionally been something created after the work in Confluence was finished: information was written and discussed in Confluence, then manually extracted and rebuilt somewhere else for a different audience. If the same content can now be transformed directly into slides, that separation starts to disappear.
AI accelerates this shift because it reduces the importance of the container. If someone can take existing knowledge and turn it into a document, a database, a whiteboard, a presentation or a summary, the format becomes secondary to the underlying context.
This may eventually make the traditional page tree less central to how we think about Confluence. The page could stop being the primary unit of knowledge and become simply one possible representation of it.
A project, for example, could begin as a Live Doc while a team is actively working on it. Some of its information could become structured data when several projects need to be compared. A whiteboard could be used to explore an idea visually. The same material could later become slides for a stakeholder presentation, while Rovo creates a shorter summary for someone who only needs the main points. Eventually, the final decisions and outcomes might be preserved as a traditional page for long-term documentation.
In that model, the same body of knowledge is no longer tied to a single format. It can move between formats depending on the task, the audience and the moment, without necessarily having to be recreated every time.
There is also a less visible consequence to this evolution. Information architecture becomes considerably more important.
Large Confluence instances already struggle with thousands of pages created over many years, often with unclear ownership, duplicated information or outdated content. Adding Live Docs, databases, whiteboards, slides and AI-generated content increases the number of ways information can exist without necessarily solving the underlying governance problem.
The challenge may therefore move from organising pages to organising knowledge itself: understanding what information exists, where its authoritative source lives, who owns it, how long it should remain relevant and how different representations of the same context relate to each other.
Confluence is gradually becoming something broader than a collection of pages.
It is starting to look more like a knowledge layer where the same information can be structured, explored, discussed, presented and consumed in different ways.
And that may be one of the most significant changes to Confluence's information architecture in years.
