Is there a way to control Rovo permissions to avoid taking agentic AI actions like create/edit ?

[Bhushan Marathe]

We have premium  JSM subscription in which we are very much looking forward to using AI. But the issue with the Atlassian AI is the Rovo chat actions cannot be controlled. We do not want users to create/edit /delete or perform any action on the issues using Rovo. 
Our intention is to use Rovo at the automation level and on the front end only to display or search for information contained within Jira and Confluence .

Even though the actions are based on user permission, we still are not confident on the actions that users might perform.

Is there any way to control it oir are there any immediate future plans to implement granular permissions for AI and Rovo?

[Matteo Vecchiato]

Hi @Bhushan Marathe ,

By the moment I don't think it is possible.

Just as idea, you could try configuring a Rovo agent with instructions about to what you allow or not. Bu anyway this don't block the other features.

Regards 

[Bhushan Marathe]

Thankyou for your response, 

I am in the process of training an agent but , I wanted to control the existing rovo chat which allows users to create ticket/request in any category.
So it is evident that currently that is not possible