Summary
Challenge: A high volume of FinOps queries via Slack has been difficult to manage. The Virtual Agent often misroutes requests or leaves key fields blank, increasing the need for human intervention and slowing ticket resolution.
Solution: A custom exception-handling Rovo Agent steps in when the Virtual Agent fails. It analyses issue descriptions, determines Query Type and Routing, updates Jira fields automatically, and adds clarifying comments using natural language prompts.
Outcome: This improves routing accuracy and speeds up resolution times. It reduces manual triage, allowing human agents to focus on higher-value work while enhancing the overall support experience.
Background
In our work with longstanding client Convex, an international speciality insurance firm, we’re continually exploring how we can automate internal processes for its business teams. Convex has embraced Rovo across its Atlassian toolset – particularly in Jira Service Management (JSM).
We have created a range of custom Rovo Agents to unlock greater time efficiencies, tighter quality control, and happier users across the organisation.
In this case study, we’re focusing on a custom Atlassian Rovo Agent to improve request triaging.
Challenge
The Convex team were handling a significant number of FinOps queries – and needed help embedding faster, and more efficient processes.
The team were already using a Virtual Agent to triage these queries. Users raised FinOps requests in a dedicated Slack channel, which were then picked up by the Virtual Agent to triage and route.
However, the Virtual Agent could not always accurately identify the right triage or routing options, which then slowed down resolution times.
Solution
To tackle this, we built a custom Rovo Agent to take over whenever the Virtual Agent cannot handle the request.
There are a lot of Query Type and Routing combinations for the queries. When users raise a request on Slack, the Virtual Agent will interact with the user to secure additional information – but if the responses from the user do not match a particular combination, the Service Agent will simply create the ticket but leave the Routing field blank.
It’s at this point that the custom Rovo Agent kicks in.
How does the request triage and routing Rovo Agent work?
We developed the Rovo Agent to act like an exception-handling agent.
If the Virtual Agent leaves ‘QueryType’ and ‘Routing’ fields is empty, an automation rule triggers the Rovo Agent to process the issue description. The new Rovo Agent will then identify the Query Type and the Routing.
Next, the automation rule will process the agent’s response and attempt to update the issue and add a comment.
Benefits
This custom Rovo Agent now takes over all failed routings. The team at Convex expects to see significant reduction in human intervention and more accurate triage/routing for FinOps queries as a result.




