Customize Rovo Chat to match company policies

[Martina_Bruehl]

Hi 
Due to the hard link between Rovo and AI features, we were forced to deactivate AI on our site because ROVO does not operate in accordance with our company guidelines. I know that access to ROVO agents can be restricted.
Is this option also available for Rovo Chat? Alternatively, is there a way to restrict Rovo Chat via system prompts or similar means, e.g., to limit the use of personal information?

[Robert DaSilva · Accepted solution]

Hi @Martina_Bruehl , welcome to the Atlassian Community!

I'm not aware of any solutions right now that would allow you to restrict Rovo usage in the way of a System Prompt. 

To better understand what your goals are, could you elaborate on how Rovo violates your internal policies? At a minimum, it may help Atlassian better understand where they need to introduce more controls for customers like you.

Cheers,

Robert

[Martina_Bruehl]

Hi @Robert DaSilva , 

We work in a Jira Service Management environment. Our service agents also have access to tickets assigned to other agents. This is necessary for day-to-day business. However, it is prohibited to create performance evaluations for others, for example. In Rovo Chat, you can do this with a simple question.

So I have the requirement to exclude personal evaluations or certain aggregations or metrics. Ideally for certain user groups.
I don't know of any way to allow access to information in Jira for a single ticket, but prevent it as a mass aggregation.

Best Regards, 

Martina

[Robert DaSilva]

Hey @Martina_Bruehl , thanks for the added information.

That sounds like a really challenging task to accomplish, for a variety of reasons.

The biggest red flag I see is that Jira Service Management (and to a certain extend the rest of the Atlassian Cloud platform) thrives on providing data, metrics, and metadata to end users. JSM has a number of built-in dashboards and reports that surface certain productivity metrics to Agents and Administrators alike. On top of this, custom reports and dashboards can be built relatively easily that can be designed to surface similar information.

There isn't really much in the way of being able to restrict the ability to aggregate data about other Agents using the built in tools, much less AI or Rovo. You could introduce Work Item Security, which would limit the ability to even see work unless a user is a participant, but this would prevent agents from assisting each other on tasks, as they won't be able to easily share the work.

I do agree that Atlassian does need to introduce better controls over how Rovo Chat and the other AI functionality behaves. I did a brief search against Atlassian's feature and bug reporting Jira instance (jira.atlassian.com/browse), but I wasn't able to find much there relating to your request.

To me, I see this more aligned with individual training, guidance, and enforcement, instead of a tools problem. If you have organizational policies that prohibit creating performance evaluations for others, that's something that should be enforced on an individual, not a tool. 

As a next step for you, I would recommend reaching out directly to Atlassian Support. You can also use the "Give Feedback" option within Rovo Chat to provide your feedback to Atlassian directly.

Robert

[Martina_Bruehl]

Hi @Robert DaSilva 

Thank you for your feedback. I will open a ticket on this. I agree with you that the philosophy of the product and the requirement don't really fit together. Nevertheless, in my opinion it should be possible to take company-specific guidelines or ethical standards into account as a basis. Or at least make it possible to use the basic AI functions without ROVO.
Thank you again for your help. 

Best Regards
Martina

[Robert DaSilva]

@Martina_Bruehl Best of luck! If Atlassian ends up creating a new feature request, I would encourage you to come back to this post and share the link. I am sure many other organizations have similar goals of adding company specific guidelines to Rovo usage, and would love to voice their support for the feature improvement.

I for one know that I would love to.

Cheers,

Robert

[Cassian Watson · Accepted solution]

This is a really valuable discussion. Being able to customize Rovo Chat so it aligns with company policies is important for maintaining consistency and compliance across teams. It’s great to see Atlassian addressing this need. I’d be very interested to learn more about how others are approaching customization—whether it’s setting guidelines for tone, automating compliance checks, or tailoring responses for different departments. Best practices and real-world examples from the community would be especially helpful.

[Rebekka Heilmann _viadee_]

I have also flagged this with Atlassian - there is certain things that my be completely fine in one setting but a forbidden practic in another. We cannot expect Atlassian to cover all bases in their system prompt but need the ability to add our own restrictions.

I'll send this discussion over to the Rovo Team as further details on the feature request.