Psst: Did you know AI can do THAT???

[#1]

Today, I want to share some customer feedback we’ve received. The first is an email I was sent from a startup customer excited to share how they’ve been able to automate many of their project management processes in Jira. How did they do it? They used AI, or to be more specific, they used Atlassian’s remote MCP server and Claude. So, with their permission, prepare to be amazed!

The email:

We spent a few hours yesterday setting up our first project in Jira using 100% Claude - and the results were AMAZING.

This is game-changing for project management, where AI-first teams were struggling to PM with good collaboration tools.

We used Claude to create epics, stories, subtasks, add story points, and more. As we complete each issue, Claude commits and logs directly into Jira. We are now implementing a smart-allocation feature that will automatically delegate issues to the right team member.

But wait, there’s more!

The customer above isn’t the only one buzzing about Atlassian’s MCP server! One of our CSM leaders received this feedback from a friend/neighbor:

"Your products are so good. We run so much of our business through Jira and Confluence, but your MCP has completely changed our workflows."

🤖 Here were some of the highlights they shared:

• Prompting Claude to summarize team priorities in Jira (key actions, tickets, etc)
• Asking for what might be missing that has an impact on the product roadmap
• Querying Intercom's mcp to prioritize feature development based on real-time customer issues
• Querying Confluence to summarize things, build new pages (or update) based on outside inputs

Cool, right? So, if you’re not sure what everyone is talking about, please read on…

MC-what-now and who is Claude?

If you’re like I was until recently, and you’re now thinking - MC-what-now and who exactly is Claude? Let me explain! First, a quick overview of Atlassian’s AI features - Atlassian products offer AI through our Atlassian Intelligence and Rovo features. And, in a nutshell, this is what each does:

Atlassian Intelligence	Rovo
Scope: Individual Atlassian cloud products, such as Jira, Confluence, Bitbucket, etc.	Scope: Rovo searches across multiple Atlassian cloud products and other third-party tools via connectors.
Features: Search for content within Jira, Confluence, etc. Generate and transform content Create automation rules Summarize Jira issues and Confluence pages, comments, and blogs Define terms in Confluence Summarize and preview Smart Links from Confluence, Jira, and Google docs	Features: Rovo Search - access your company's data across Atlassian and connected 3rd party apps Rovo Chat - conversational interface for tasks like finding answers, generating ideas, etc. Rovo Agents - configurable AI teammates to collaborate and move work forward Studio App - create everything from AI agents to automation and apps in one unified experience

So, hopefully that clarifies Atlassian Intelligence vs Rovo, so, what’s this MCP server thing?

MCP (Model Context Protocol) is an open protocol that standardizes how applications provide context to AI large language models or LLMs. Think of MCP like a USB-C for AI applications. Just as USB-C provides a standardized port to connect and charge your stuff, MCP provides a standardized way to connect AI models to different data sources and tools.

So, Atlassian’s MCP server gives you a way to connect Atlassian’s data and AI to other AI tools, such as Anthropic’s Claude. And, what do we get if we connect these different AI tools?

NO, NOT Skynet and terminators!

We get the ability to access information from Jira and Confluence wherever you use your other AI tools (right now, that’s Claude). This lets you do all sorts of things like:

• Summarize Jira work items or Confluence pages
• Create Jira work items or Confluence pages directly from Claude
• Get the model to take multiple actions in one go, like creating issues or pages in bulk
• Enrich Jira work items with context from many different sources that Claude has access to
• And so much more!

Wanna learn more?

Here’s some great resources on Rovo and Atlassian’s MCP server so you can learn more.

Atlassian Intelligence:

• What is Atlassian Intelligence?

• Atlassian Intelligence Trust FAQ

• Overview of Atlassian Intelligence features

• Guide: Get started with Atlassian Intelligence

Rovo:

• What is Rovo?

• Rovo data, privacy and usage guidelines

• Rovo Connectors

• Rovo Resources

• 3 Rovo Agents That Make Your Life Easier

MCP server:

• Connect AI to your enterprise knowledge with Atlassian’s MCP server

• Introducing Atlassian’s Remote Model Context Protocol (MCP) Server

• Using the Atlassian Remote MCP Server beta

• Form for MCP Server bugs and feedback

In the comments let us know how you plan to use AI, Rovo or MCP!

[Gary Spross]

Here's a question I've been unable to answer... If we have Claude, and it's connected to Atlassian via the MCP server, what use is there for Rovo? 

Search: Claude can be connected to many tools, just like Rovo, and search across them

Chat: Claude is an entirely conversational interface

Agents: Perhaps this is the one area where Rovo has the advantage?

[Peggy Graham]

I can't speak to Claude, so please let me know if Claude does this too, but I was thinking the Rovo agents are the differentiator?  Via Claude, you can ask it to have a Rovo agent do something in Atlassian or any other connected tool.

If nothing else, well, you have options and can use either to find the most efficient steps to accomplish what you need to do.  :-) 

[Gary Spross]

In a perfect world, we would have all the tools for sure. However, cost, although not enforced yet for Rovo, is the concern. Duplication of tools is an unnecessary expense, so the individual tools need to be justified. 

Since we utilize Claude, what I would like to have, once Rovo usage limits are enforced, is the ability to limit how users can utilize Rovo. Since Claude can do the Rovo Search & Chat features, I would only allow users the ability to utilize the Rovo Agents feature. This would ensure that any consumption is related to the aspect of the tooling that isn't available elsewhere.

[Peggy Graham]

Ooooohhhhhh, thank you, I understand your meaning now!

There is a feature request for this, please see: ROVO-107 - Enhanced Granular Control Over AI Features.  Please take a look at that and upvote if you agree.

NOT a part of what you asked, but while I was checking on that one, I found a related request that might also be helpful: ROVO-23 - Ability to assign Rovo access based on user group membership.

I hope that helps!

[Gary Spross]

Both of those suggestions are exactly what I'm be looking for, thank you!

Hopefully having them listed on this post will get more eyes on them and therefore more votes because both of these features are highly valuable and highly necessary.