Turn User Comments into Live Actions: Using ROVO magic in Jira workflow

[Raghavendran_Narayanan]

Ever wished Jira could respond like Siri? You say, “Hey Siri, play my favorite song,” and boom—it happens.
Now imagine this in Jira: You type “Hey Rovo, break this feature into stories” … and a minute later, stories are created automatically.
Sounds like magic? It’s real. Let’s dive in.

Use case :

With ROVO, a simple comment in Jira can trigger intelligent automation—creating stories, refining features, and more.

Here’s the secret sauce:

• Step 1: You comment on a Jira feature starting with “Hey Rovo”
• Step 2: ROVO detects your intent (create stories, refine, suggest titles, etc.)
• Step 3: Jira automation kicks in, calls the right ROVO agent, and… Boom! Work happens.

Disclaimer : This is an advanced topic and i am not going to go through every step of implementation. However, i will guide you in high level about how to get this done.

What You Need

• 2 ROVO Agents
  • Intent Detection Agent: Understands what user wants
  • Story Creation Agent: Builds stories based on your feature
• 2 Jira Automations
  • Comment Monitoring: Triggers on “Hey Rovo”
  • Story Creation Trigger: Fires when intent is confirmed

The Magic Explained

• Intent Detection Agent: Reads your comment, figures out your intent, and responds with an acknowledgment + label (e.g., “CreateStories”).
• Story Creation Agent: Takes the feature URL + instructions, generates stories with titles, descriptions, and acceptance criteria.
• Automation Flow:
  1. Comment → Intent detected → Label added
  2. Label triggers story creation → Stories appear → Summary posted

Still curious? Here’s the Blueprint to make it happen.

1st ROVO agent :

• Name : Story creation agent
• Input : Feature URL and instruction to create user stories
• Process: Reviews the feature, understands requirements, creates stories
  (Tip: Enhance later with duplicate detection, validations, etc.)
• Output: A clean JSON string with wiki-friendly content
  (Sample JSON template at the end)

2nd ROVO agent :

• Purpose: Understand what the user wants
• Input: Free-text comment (e.g., “Hey Rovo, create stories”)
• Process: Semantic analysis → maps intent to a label
  (Examples: “CreateStories”, “RefineMe”)
• Output: JSON with:
  • Acknowledgment message (or error if intent unknown)
  • Intent label (empty if unknown)

1st Jira Automation : "Comment monitoring"

•  Trigger : Work item commented , "comment is the main action" . Conditions : comment lower case contains the expression. (?i)^\s*hey\s+rovo\b
• Call the Intent Detection ROVO (2nd ROVO agent) with the user comment as input for the agent
• Process the output JSON through the variable : {{agentResponse}} . 
• If the label exist, set the label at the item where the comment was made.
• Edit the comment (Triggered comment) with ROVO acknowledgement message, as  {{agentResponse.asObject.ReturnMessage}}

2nd Jira Automation : "Create stories agent invoking through Labels"

•  Trigger : Field value changed. Condition : Labels added and Label contains : CreateStories
• Invoke the 1st ROVO agent : Story creation agent with the input of feature URL {{issue.url}} and any special instruction for user story creation
• Process the output JSON through the variable : {{agentResponse}} . 
• Retrieve the stories through : {{agentResponse.asObject.stories}} assign to variable story , in the For Each
  • Then use Create Stories under the For each and populate the title : {{story.title}} , description : {{story.description}} , AC 

• Retrieve the stories through : {{agentResponse.asObject.stories}} assign to variable story 
• Finally outside of For each, give a Feature level comment on summary of stories created. Through {{agentResponse.asObject.summary.wikiTable}}.
• Be sure to check the box for  : 
  Check to allow other rule actions to trigger this rule. Only enable this if you need this rule to execute in response to another rule. since this automation should be triggered from 1st automation output.

Congratulations! You’ve just made Jira conversational. Next time you think, “I wish Jira could do this for me,” just say:

“Hey Rovo…” and let the magic happen.

JSON output template for 1st ROVO agent (Story Creation agent)

{
"mode": "automate",
"parentKey": "<Feature/Epic Key>",
"board": "<Board Key>",
"stories": [
{
"title": "<Concise, actionable story title>",
"description": "As a <role>, I want <goal>, so that <benefit>....
"acceptanceCriteriaB": <Generated AC of ROVO>
}
// ... more stories as needed
],
"summary": {
"rows": [
{
"requirement": "<Requirement text>",
"coveredBy": "<Story title or child issue ID>",
"status": "<Newly generated/Skipped new story creation>"
}
// ... more rows as needed
]
}
}