nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Confluence Connector for power platform Tags

Daniel Gray

I'm working with the Confluence Connector for power platform and it seems useful. We are looking at using the connector to surface the data in our confluence to Gen-AI (Copilot etc).

However, we have a lot of older out of date information in our Confluence sites. The connector seems to just pick all pages (unless they are archived).

Is there any functionality (current or planned) where maybe the connector could be configured to pickup tags?

I'm thinking that we could add a tag to pages iteratively & then the connector is configured to just pick pages with a certain tag. This means we could easily control the pages that the Gen-AI agent sees to ensure out of data information isnt surfaced this way.

Accepted answers

Germán Morales

Hi @Daniel Gray,

The behaviour you describe, the connector pulling every published page and skipping only archived ones, is the documented default of the Microsoft 365 Copilot connector for Confluence Cloud. The useful part is that what you want already exists. That connector lets you scope what gets indexed with a Confluence Query Language (CQL) filter, and CQL filters on labels, so your tag based approach works today rather than being a roadmap request.

Add a label to the pages you want surfaced (Confluence calls tags "labels"), then set the connector's CQL filter to index only those, for example:

type = page AND label = "approved"

You can narrow it further, such as space = "DOCS" AND label in ("approved", "current"), and the setup includes a Preview results step to confirm the page set before you sync. The filter lives in the connector's content scope on the Microsoft side, while the label and CQL syntax are Atlassian's, under Advanced searching with CQL. If you turn out to be on the Power Automate or Logic Apps Confluence connector instead, the same CQL idea applies through its content actions, but for feeding Copilot directly the Copilot connector is the natural fit.

All comments

Germán Morales

Hi @Daniel Gray,

Add a label to the pages you want surfaced (Confluence calls tags "labels"), then set the connector's CQL filter to index only those, for example:

type = page AND label = "approved"

Tomislav Tobijas

This is a Rovo limitation as well (well, all AIs it would seem): ROVO-471: Add Label-Based Filtering for Confluence Knowledge Sources in Rovo

It might be the case that, when they roll this out for Rovo, other connectors/integrations will support this functionality as well 👀

Daniel Gray

Thanks for the responses.

In case someone else find this when searching; if you do the connection straight in CoPilot studio the CQL filtering isn't there, but it appears that CQL is an option if the connection is setup in 365 admin centre (https://learn.microsoft.com/en-us/microsoft-365/copilot/connectors/deployment-overview).

Welcome!

It looks like you're new here. Sign in or register to get started.

Quick Links

Find more posts tagged with

confluence

cloud

confluence-cloud