Accelerating Customer Support with Rovo Agents in Jira Service Management & Confluence

[#1]

🤖 Rovo Support Agent (Level 1): Assisted Ticket Handling

 1. Ticket Creation

 A customer submits a request via the JSM portal, email, or chat—the usual entry points.

 

2. Smart Manual Trigger

The support engineer or dispatcher clicks the “Draft First Response” button, manually activating the Support Agent.Instantly, Rovo analyzes the ticket summary, labels, and stack traces. It searches historical tickets and Confluence articles, then ranks similar results by confidence and relevance.

 3. Internal Comment Draft

 The agent responds internally with a suggestion containing:

 A concise diagnosis summary

• Top 3 related KB links or resolved ticket IDs

• A human-like draft reply, trimmed to 1–2 paragraphs
  

 4. Engineer Review & Response

 The engineer reviews, edits if necessary, and sends the polished response. It’s quick, accurate, and reduces repetitive work.

---

📄 Rovo Confluence Article Drafter (Level 2): Knowledge Base Automation

 5. Article Generation 

The agent pulls all relevant information—ticket description, attachments, public comments, and the root cause field—then builds a draft Confluence page using your predefined template, including:

• 🧩 Problem Statement

• 🖥️ Environment Details

• 🔧 Steps to Fix

• 🧾 Customer-Safe Summary

6. Human Review & Publish

A support lead is to review, refine, and publish the draft. The new article is then linked back to the original ticket, strengthening your self-service ecosystem. Use labels for the published contents those can be later filtered in confluence via Scheduled Automation for review and publish.

• 

7. Pro Tip💡

To get the most accurate responses from your Rovo agents, feel free to customize your prompt instructions using Atlassian’s guidance—like the “Example instructions” in their support docs . Think of this as fine-tuning your AI sidekick’s behavior: you may need a few rounds of tweaks and trial-and-error to get it just right.

https://support.atlassian.com/rovo/docs/write-instructions-for-your-agent/

That’s also the reason I use a manual trigger for the first agent rather than fully automating it—from the “Draft First Response” button. This human checkpoint lets your team review and refine the agent’s output before sending it, ensuring the AI meets your accuracy standards and tone preferences.