Introduction
Knowledge collections are the central place to curate and manage your organization's knowledge inside Docebo. A collection is where you decide, piece by piece, what knowledge from your connected systems should actually be made available to your users, and it's the only way to make that knowledge accessible on the platform.
Content comes from connectors, not uploads
Knowledge collections don't support uploading files directly. Content can only be added to a collection from a system that's already connected through a Knowledge connector, which means at least one connector has to be set up before you can build your first collection.
This is a deliberate choice, not a current limitation. Uploaded files are static: the moment a file is uploaded, it stops reflecting what's actually in the source system, and someone has to remember to re-upload it every time it changes, and to remove it when it's retired. It would also mean managing a separate set of permissions for that file inside Docebo, on top of whatever permissions it already has at the source. Building collections exclusively from connected systems avoids both problems: content stays linked to its source, so there's no manual lifecycle to maintain, and permissions come from the source system, so there's no separate access logic to keep in sync.
For details on setting up a connector, see Knowledge connectors.
Managing your collections
Superadmins manage collections from the Knowledge collections overview, found under Content and delivery > Knowledge collections. This overview lists every collection that exists, split into those that are published and those still in draft, and it's where you create a new collection or open an existing one to edit it.
Collection settings
Every collection has a settings area where you define its base information: name, icon, and description. This is also where you set the collection's audience, restricting its visibility to specific branches or groups.
Audience is an additional access layer on top of, not a replacement for, source system permissions. Restricting a collection to a branch or group only ever narrows down who can access it inside Docebo; it never grants access to knowledge a user isn't already permitted to see in the source system. Source permissions are always enforced first, regardless of how a collection's audience is configured.
Adding content
A collection is organized into sections, and you add content to each section by selecting it from a connected system. What you select depends on the type of connector.
Indexing connectors
The file browser mirrors the folder or space hierarchy exactly as it exists at the source, and lets you select content at any level of that hierarchy.
Rather than adding individual items one by one, you can also select an entire folder or space. Doing so has a real advantage: if new content is later added inside that folder or space at the source, it automatically becomes part of the collection too, with no need to go back and add it manually. This is worth keeping in mind when deciding how granular to be when building a collection: selecting at the folder or space level keeps a collection current with minimal upkeep, while selecting individual items gives tighter control over exactly what's included.
When you add content, it's extracted and indexed, which counts toward your indexed data volume. Indexed data volume is billed according to your contract. For details about included volume and pricing, contact your Docebo representative or reach out to Docebo using your platform's Help center.
Live connectors
For live connectors, you add a Project (Jira), a Channel (Slack) or an Object (Salesforce) to the collection instead of browsing individual content. Nothing is copied or indexed, so adding them never uses any data volume.
When a user's prompt searches the collection, Docebo searches the added projects and channels at that moment, using the user's own account. New activity in them is therefore included automatically, and users only see what they could open in Jira, Slack or Salesforce themselves.
Publishing
A collection stays in draft, visible only to Superadmins, until you're ready. Once you've finished adding content, publishing the collection is what makes it accessible to users, subject to the source system permissions and any audience restriction described above.
Where published knowledge shows up
Once a collection is published, its content can surface for users through the AI assistant and agents. See the Docebo Knowledge article for a full rundown of how each of those surfaces works.