Why Is Rovo Usage Showing “Jira Administration”? [Champions Slack Insider]

[#1]

One of the best parts of the Champions Slack is the level of questions that get asked—the kind that don’t always make it to the public forum but should, because they help everyone.

So I’m starting a new series to bring those conversations forward.

Let’s start with a great one from @Fazila Ashraf 

Hi all,

I’ve been diving into our org's Rovo usage and credit consumption reports and noticed a trend I’d love to sense-check with you.

We are seeing fantastic organic growth with Rovo. We had plans to push adoption further, but our recent analysis shows that at this rate, we’ll exhaust our credits well before our billing cycle While Atlassian hasn’t enforced consumption limits yet, we want to avoid a "bait and switch" scenario where we encourage usage only to restrict it later due to budget caps.

Our consumption reports show a significant amount of credits attributed to the "Jira Administration" category. However, these are being triggered by power users who do not have Jira Admin permissions.

I suspect this might be related to the Rovo portal agent or specific "Actions" the agent is taking that are being classified as administrative on the backend.
Has anyone else seen this "category creep"? Is there a specific type of Rovo query or Agent action that triggers the "Administration" classification? Any insights on how to better forecast this would be huge.

What’s Actually Happening

This is a great example of how Rovo reporting doesn’t always map cleanly to user roles.

The key insight: “Jira Administration” reflects the type of action or data accessed, not the user’s permission level

In practice, this means:

• Actions involving configuration data (workflows, schemas, metadata)
• Queries that touch system-level structures
• Certain agent-driven actions or skills

…can all be categorized as “Jira Administration” on the backend—even if the user is not an admin.

Where This Gets Confusing

Right now, there are a few gaps that make this harder to interpret:

1. Reporting is more detailed in exports than in the UI

• The CSV export (Platform Usage → Rovo credits) shows deeper categorization
• The UI views don’t always surface this level of detail

2. Attribution isn’t fully transparent yet

• It’s not always clear which specific action triggered the classification
• Agents and portal interactions may abstract what’s happening behind the scenes

3. Beta features may affect categorization

• Participation in programs like Rovo for Admin can influence what shows up
• Not all environments will see the same categories

Why This Matters (Especially Right Now)

Even though Atlassian hasn’t enforced consumption limits yet, this is exactly the kind of pattern teams should be paying attention to.

Because eventually:

• Credits will be enforced
• Usage patterns will matter
• And teams may need to adjust behavior quickly

The last thing you want is to encourage adoption—only to scale it back later due to cost surprises.

What You Can Do Today

Until reporting becomes more transparent, here are a few practical steps:

1. Monitor patterns, not just totals

Look for:

• Spikes tied to specific teams or use cases
• Repeated agent interactions
• Queries that touch large datasets or system-level data

2. Be intentional with high-cost use cases

Not all AI usage is equal.

Watch for:

• Broad queries across large Jira datasets
• Agent automations that trigger multiple actions
• Repeated exploratory queries

3. Use exports to get deeper insight

The CSV export from Platform Usage → Rovo credits currently provides the most detailed breakdown available.

4. Validate assumptions with Atlassian

If something doesn’t make sense:

• Raise a support ticket
• Loop in your Atlassian account team
• Share patterns (not just examples)

This helps improve both clarity and the product itself.

Right now, forecasting Rovo usage isn’t an exact science. The best approach is:

Observe → identify patterns → adjust usage intentionally

Over time, as:

• reporting improves
• governance features mature
• and pricing models stabilize

…forecasting will become more predictable.