Key takeaways
- A single Atlassian catalogue entry covers Jira and Confluence; the catalogue may change.
- Each person signs in with their Atlassian account, so the assistant sees only what that person can see.
- Launch work is the strongest use: spec in Confluence, brief in chat, tasks in Jira.
- Reading a page and filing a ticket are both MCP calls under the execution mode.
Does one connection cover Jira and Confluence
Yes. The Pebbles Ai MCP catalogue lists Atlassian once, for your Jira and Confluence data and actions, so a single connection reaches both. The tools come from Atlassian's own server and may differ between the two products.
Any user can connect it from Settings, Connections, with their Atlassian sign-in. Page and project permissions stay Atlassian's: a space you cannot open in Confluence stays closed to your assistant too.
How do you get a Confluence AI summary of a launch spec
Ask the Marketing assistant to read the spec page and summarise it for a go-to-market audience rather than for engineers. Name the page, approve the read if you work in Manual mode, and ask for what changes for customers, who it is for and what could confuse them.
The assistant writes from Company Intel as well as the spec, so the summary uses your offerings, audience and messaging rather than the internal feature name. That is the difference between a summary and a usable brief.
How do you go from spec to go-to-market brief
Ask for the brief in a fixed outline, in the same chat as the summary. An illustrative outline a product marketing manager might use:
- The change in one sentence a customer would understand.
- Who it matters to, by role and segment.
- Three key messages, written in the brand voice.
- Assets needed: announcement email, help article, sales talking points, social post.
- Open questions for product before launch day.
How do you file launch follow-ups as Jira tickets
Turn the assets list into tickets by asking the assistant to file one Jira issue per asset in the project you name. Each filing is an MCP call: in Manual mode you approve each card, in Plan mode they are queued, and in Auto mode they run without asking.
Keep Jira's workflow, boards and assignment rules as they are. The assistant creates the ticket; your team still runs it in Jira.
What stays outside the assistants' reach
Confluence pages are not copied into Company Intel or memory by being read. If a spec should inform every future chat, an admin adds it to Company Intel as an uploaded file or pasted text.
The Guider does not use MCP tools, and Atlassian cannot be added any way other than from the catalogue in Settings.
How to run launch work across Confluence and Jira
One chat carries the whole job, so the brief and the tickets stay consistent.
- Connect Atlassian from the MCP servers group in Settings, Connections, signing in with your Atlassian account.
- Ask the Marketing assistant to read the launch spec in Confluence and, in Manual mode, approve the read.
- Ask for a customer-facing summary, then a go-to-market brief in your agreed outline.
- Review the brief and ask for one Jira ticket per launch asset in the project you name.
- Approve each Jira call, then share the brief with the launch team.
Where Pebbles Ai fits
- FoundationIntegrationsAtlassian, for Jira and Confluence, is in the Pebbles Ai MCP catalogue (the catalogue may change), connected per person with Atlassian's sign-in.
- FoundationCompany IntelCompany Intel holds your offerings, audience and brand voice, so a brief drafted from a spec speaks the way your company does. Admins add files or pasted text to it.
Common questions
Do I need separate Jira and Confluence connections?
No. One Atlassian connection from the catalogue covers both, with tools supplied by Atlassian's server.
Can the assistant see Confluence spaces I cannot open?
No. You sign in with your own Atlassian account, so the assistant works within your existing permissions.
Does reading a Confluence page add it to Company Intel?
No. To make a document part of Company Intel, an admin uploads it or pastes its text there.
About this page. Written by the Pebbles Ai team for people who run sales, marketing and growth. Last updated .
