How I Built a Rovo Agent That Became My Sprint Buddy!

[#1]

Sometimes the best ideas are born out of frustration. A few weeks ago, I was staring at my Jira board late in the evening, trying to prepare for the next day’s sprint review. The board was packed, the filters were messy, and I found myself asking:

“Why am I spending so much time just answering basic questions about our sprint?”

That’s when it hit me—I didn’t need to. I could teach a Rovo agent to do it for me.

The Real-World Challenge

In my team, sprint reviews often start with three key questions:

1. What progress have we made since the last review?
   
   
2. What blockers are still unresolved?
   
   
3. Are there any risks for the upcoming release?
   
   

Answering these questions usually took me 30–40 minutes of running filters, exporting reports, and summarizing in slides. 

So, my goal was simple: build a Rovo agent that could act as my “Sprint Buddy” and deliver these answers instantly.

Building My “Sprint Buddy” Agent

Here’s how the journey unfolded:

1. Open Studio

Go to Confluence or Jira → Open Studio → Click Create Agent.

2. Name Your Agent

• Example: “Sprint Buddy”
  
  
• Description: “Summarizes sprint progress, blockers, and risks for reviews.”

3. Add Agent Instructions

This is where the magic happens. I wrote down what I wanted the agent to do, like I was giving a teammate a job description:

Agent Instructions:

At any point during an active sprint:

1. Summarize sprint progress (completed vs remaining issues).
   
   
2. List unresolved blockers with issue keys and assignees.
   
   
3. Highlight overdue tasks and explain potential risks.
   
   
4. Present the output in a clear bullet-point summary.
   
   
5. Use emojis where helpful: ✅ for done, 🚩 for blockers, ⏳ for overdue.
   
   

4. Add Conversational Starters

These help the team ask the right questions:

• “What’s the current sprint progress?”
  
  
• “Which blockers are still unresolved?”
  
  
• “What overdue issues should we watch out for?”
  
  

5. Add Knowledge

Connect your Jira project data. I limited scope to:

• Issues in the active sprint
  
  
• Blockers flagged as Priority = Blocker
  
  
• Tasks with due dates < today
  
  

 Tip: Don’t give the agent access to everything. Keep it focused so results stay relevant.

6. Add Actions

I set it up so the output could be shared in Slack or saved to a Confluence page. That way, updates weren’t just for me—they became visible to the entire team.

7. Test & Refine

The first run was messy—it returned raw issue IDs and too much detail. After refining prompts, I got a clean, readable summary like this:

✅ 5 issues completed, 8 in progress
🚩 2 unresolved blockers (ABC-123, DEF-456)
⏳ 1 overdue story (GHI-789, 3 days late)

Perfect for dropping straight into Slack before sprint review.

Lessons Learned Along the Way

• Think like a teacher: Your prompts are lesson plans. The clearer you are, the better the agent performs.
  
  
• Don’t chase perfection: Even an early version saves time—improve it iteratively.
  
  
• Small wins matter: This simple agent inspired teammates to explore their own use cases.
  
  

Why This Matters

This wasn’t just about saving myself 30 minutes. It was about showing my team what’s possible with Rovo. Now, teammates are building their own ideas:

• QA: an agent that flags untested requirements.
  
  
• Product Owner: a release readiness checker.
  
  
• Developer: an agent that drafts release notes.
  
  

Your Turn

If you’re thinking about building your first agent, here’s my advice:

1. Start with a real pain point in your workflow.
   
   
2. Keep the scope small—solve one problem well.
   
   
3. Share your agent early, even if it’s rough.
   
   

Don’t underestimate how inspiring a small automation can be for your team. 

#AtlassianChampions #Rovo #JiraAutomation #AIinJira