I've been seeing more and more clients getting curious about AI and looking for ways to use it to save time in their daily work and this solution seemed like a good one to share here at the community.
Recently, I built a JSM automation that generates a Postmortem page whenever a Highest priority incident is resolved.
The workflow is straightforward:
The rule is triggered when an Incident transitions to Resolved.
It validates the incident priority.
A Rovo Agent is invoked to analyze the incident and generate the postmortem content based on the issue data.
Finally, Jira Automation creates a Confluence page containing both the incident metadata and the AI-generated analysis.
The generated page includes information such as:
Incident summary
Reporter and assignee
Timeline
Description
Resolution
AI-generated root cause analysis, impact assessment, lessons learned, and recommendations
The nice part is that the Rovo Agent isn't just summarizing the ticket; it analyzes the issue context and comments as well. Even cooler, if it's connected to your KB, so it can use information to provide additional context.
For me, this solves two common problems:
Postmortems are often skipped because they're time-consuming.
When they are written, the format and level of detail usually vary between engineers.
I was also impressed by how easy it was to put together this combination of Rovo + Automation. It opens up some really interesting possibilities for combining workflows with contextual AI, without the need to build a custom integration.
Here's an example using HP references, so we don't bore ourselves to death reading a regular postmortem. 😄
I know there are plenty of opportunities to expand this workflow; for example, automatically creating Problem tickets, notifying stakeholders, or enriching the analysis with additional sources. I'd love to hear what other use cases you've explored.
Ana Vitória Selista
2 comments