Rovo as Jira Admin "Sidekick:"?

[#1]

Initially opened as a question by mistake (see here).

I was wondering if (and when) will Atlassian take AI, Rovo or a new one) to the next step of assisting Jira admins with the heavy lifting of Jira administration.

Jira administration is mostly a click-ops mess. Most definitely when there is no or little central management of the the tool, and when Team spaces are involved.

The existing API for administrating Jira, is complex, full of inconsistencies, documentation errors, and even errors on how the API behaves, thus using the API for limiting down click-ops is also frustrating, more over if one is looking for a generic solution which can support a IaaC approach with proper CI/CD to modify Jira's configuration (personally working on such python package - internal to the company).

Doesn't it make sense to have Rovo also broaden to support administration tasks (adding fields/contexts, creating workflows, updating screens, etc,) based on well provided prompt? 

Obviously this will need to be supported first in a sandbox with the ability to promote what ever Rovo did, once reviewed, to production. which too should be supported by Rovo. 

What do you think?

[Matt Smith]

Given that Jira is allegedly an ITSM tool,  and yet it does not provide even a single Sandbox for standard tier, I think you are clutching at straws for this to happen ;-)

[s_gridnevskii]

I think it is too dangerous to allow AI to make changes in config because some of them are irreversible. E.g. if a new workflow does not have some steps and you decide to rollback you will need to check history in each issue to find out what was the original status, make a temporary link in old workflow, transit issue and rollback old workflow. This is not what I will allow ai to do.

[Jimmy Seddon]

Hey @Shai Gilboa!

While I can appreciate where your question is coming from, I think you also explained one of the biggest challenges Atlassian has on delivering this ask.  If there are inconsistencies in the Admin APIs then Rovo is likely to struggle to make the same changes just as much as we do.

Also, I believe their focus right now is getting people to adopt Rovo and adding functionality that only benefits us admins isn't going to be what gets C-Suite at companies on board. 

[Urmo]

I believe that today AI should remain more of a supporting tool, rather than something that independently and aggressively makes changes, modifies things, and then later requires debugging to understand why something broke.

Rovo is nice, but at times it is already limited by the fact that when you ask questions related to Jira Cloud, it sometimes provides documentation that actually applies to Data Center (DC), which can be misleading. Of course, the issue may also be with how the question is asked — AI may not yet fully understand the context — and the documentation itself can be problematic. If you are not very familiar with Atlassian, some automations and configurations may remain incomplete because field value names don’t match the documentation, or the documentation has not yet caught up with how things actually work.

I see Rovo as something that could take care of background housekeeping tasks. Over the years, so-called “operational clutter” accumulates that people don’t have time to clean up themselves, and Rovo could summarize or identify these things in the background. For example, a custom field that is not used in any project, including archived ones. Or custom field values that have never been used — asking whether they should be deleted or archived.

Rovo could also assist with building dashboards.

[Jean Gordon]

If this ever changes please do NOT make it mandatory as we already have issues with many other things with tickets outstanding.

[Dave Meyer]

It's interesting to see the comments here. @Shai Gilboa this is an excellent question and very well aligned with how we are thinking about Rovo and Jira administration. Short answer – we are already investing quite a bit in this and expect to launch a beta program for more Rovo features for Jira admins early in early 2026. As you might have seen, we've already launched Rovo-assisted workflow creation and editing in the workflow editor. We are also making investments in general "sidekick" type functionality and specifically in common high-complexity admin tasks like resolving permissions problems.

I think Rovo is going to be transformative for a lot of Jira admins:

• For beginners and inexperienced admins, it will dramatically reduce the learning curve. More advanced configuration will become much more achievable without first building an encyclopedic knowledge.
• Advanced admins will become far more productive. Rovo will make it far more trivial to execute tasks in bulk or through automation by removing the bottleneck of generating scripts, and it will help suggest the right cleanup or maintenance actions to take.
  • Obviously, we will invest in specific Rovo capabilities to make these actions more powerful and reliable, but I assume that Jira admins today are already using AI tools to help them script common tasks.

Every experienced Jira admin I've met has a story of an inherited Jira instance that had a chaotic mess of configuration that they've had to clean up. Rovo has the opportunity to make that type of project much faster and easier, or prevent it from being necessary in the first place.

To address a few of the comments here:

1. I think we're still a ways off from Rovo becoming proactive for admin/config changes - everything will still require a human in the loop to approve a specific set of deterministic actions.
2. Developing additional Rovo capabilities for Jira admin will also require us to address some missing Jira REST APIs for admins, I think this will end up being a double win.
3. We're definitely aware of Rovo mixing up Cloud and DC documentation or relying on outdated documentation – this is something we are specifically working to address while still benefitting from the incredible corpus of good practices and tips/tricks on the internet that isn't in Atlassian's docs.

@Jimmy Seddon you make a reasonable point about Rovo adoption. But I would argue that a great way for Atlassian to increase the number of people using Rovo is to increase the number of people using Jira. If we can make Jira admins out there more productive and they can spend less time doing simple, rote tasks, they can spend more time helping support good practices and driving adoption within their organizations. That's the best possible outcome for our customers, and if more people are able to get value from Jira and Rovo, then Atlassian will also do well in the long run.

[Devlend Maul]

Not saying they will or won't get to that point but they are starting to crawl in that direction with their "Rovo Skills for Jira Admins" ERP.

https://community.atlassian.com/forums/Jira-Cloud-Admins-articles/Rovo-Skills-for-Jira-Admins-early-access-program-Application/ba-p/3161460

[Haddon Fisher]

I've been using Claude to handle some tasks and been really enjoying it. There's a class of work that is conceptually pretty simple to understand but hard to execute for some reason or another which it can handle extremely well - bulk archiving issues, for example.

However, I have to admit that most of the use-cases I can think off the top of my head are not really arguments for AI but more better functionality. The reason I needed to use Claude for that was because it's not possible to do this using "bulk update" or any other UI tool.