How are You Using Rovo Agents? [Earn Badge + 100 Kudos]

[#1]

Welcome to Week 2 of Go AI-Native with Rovo! This week, it’s all about agents.

Whether you’ve built a custom agent tailored to your team’s workflow or you’ve found a go-to agent that’s changed how you work, we want to hear about it!

The real value of this next challenge is seeing how fellow Community members are using agents as AI teammates. Maybe you've built one that handles weekly reporting or streamlines your team's intake process in Jira - chances are someone else in the Community is trying to solve the same thing.

⭐️ This week’s challenge

Share a screenshot in the comments below showing us:

1. An agent you created – tell us what it does and the problem it solves, or

2. Your favorite agent – and why it's earned a spot in your workflow

Drop your screenshot, tell us the story behind it, and you’ll unlock:

• 🏆 Agent Architect badge

• 💰 100 Kudos points

New to agents? Start with this learning course to build your first one. Once you've got the basics down, here are a few ideas to try:

• A competitive agent that tracks competitor updates so you're always in the loop

• A meeting prep agent that gathers context, action items, and docs before your next call

• An onboarding agent that answers common new-starter questions from your team's Confluence spaces
  
  

💡Insider Rovo Tip: When you write an agent's instructions, you're not just writing a prompt – you're programming its decision logic. You can include conditional behavior ("If the user asks about pricing, always check the Q4 pricing page first"), prioritization rules, and output formatting. The more specific your instructions, the more deterministic the agent becomes.

Share your agent below and and take a scroll through to see what others have posted. You just might find your next favorite workflow. 👇

[Sami Shaik]

One front door for every internal service request: our Rovo Service Agent

Our staff raise IT, HR, facilities and finance requests across many Jira Service Management portals, each with its own request types and Proforma forms, some over 50 questions long with conditional sections, cascading categories and mandatory attachments. Nobody remembers where anything lives, so tickets landed half-filled or in the wrong queue.

So we built a Rovo agent that turns all of it into one conversation: "I need to request my travel allowance" and the agent finds the right request type across every portal, then walks the user through it question by question, with a live progress bar (▰▰▰▱▱ 33% · Question 5 of ~12) so long forms never feel endless.

The story behind it, and our biggest lesson: our first version let the agent handle the form fields itself, and on complex forms it would occasionally invent field IDs or mis-map answers. The fix was an architecture decision that changed everything: make the agent a thin relay. The conversational layer lives in Rovo Studio, while a Forge app owns all field logic. It tracks which question is next, validates every answer server-side, resolves people and option IDs, evaluates conditional visibility, and builds the exact JSM payload. The agent converses, the app guarantees correctness. Since then: no hallucinated fields, ever.

Favourite recent win: chat can't carry file attachments, so the agent hands users a small Forge page that already holds their answers. They drop the file, click submit, and the request is created with the form completed via the Forms REST API. Users come back to chat and the agent already knows their ticket number.

If you're building an agent on complex forms: keep the instruction short, and push every rule you can into actions. Your agent gets smarter the less it has to remember.

[Jason Krewson]

I've created a couple really helpful ones so I will list them all below, but the one I want to highlight would be the Jira & Confluence Guide. This agent answers Jira and Confluence questions by guiding users to the most helpful documentation in the <<Company Name Here>> Atlassian Confluence space and offical Atlassian Documentation. This was the first Rovo agent I created and it helps users navigate all our our documentation around Jira and Confluence making it easier to find what they are looking for.

[Ross LeMere]

While simple, I created a Time Tracking Reminder Rovo Agent to assist our IC's when it came to tracking their time. In the industry my company works in, tracking time towards a contract is very important. We also have gone through multiple changes in process and how we track time and noticed that it was causing some thrash amongst our team members. In order to assist them and make their lives as easy as possible when it came to tracking their time, this Rovo Agent was created.

The screenshot below show an example of the Agent providing Time Tracking best practices, specifically aimed at new employees that might not be familiar with time tracking to the degree that we do. You can also ask this Agent to remind you during the day to track your time, explain the differences between our two apps that require time tracking and how to keep them in sync, etc. all with a friendly and helpful tone because no one likes to track their time and we want to keep it low stress.

[Tomislav Tobijas]

Some really cool examples here :)

I've got a fun one, but I'll share that in a separate article I'm preparing. 😉

As for the one I can share here, a couple of months ago, I built an agent that works with our (relatively new) process for estimations. The agent analyzes the requirements and then, based on previous projects, creates a structured Confluence page with some baseline estimations based on current knowledge.

So far it has worked really well, but we're already fine-tuning the process and building a new version of this (potentially even with Forge) to be able to handle more complex scenarios.

I can just say, things look really promising 🤞

[Amanda Barber]

I've been working on an agent that can take our current user guide structure/formatting (from an external product) and start a new user guide based on notes and a Loom for a new feature. I haven't quite perfected it, but this thread is giving me the nudge to work on improving it again!

[Carsten Severin _KUMAVISION AG_]

I’ve already built a few agents, but the one I personally find most interesting is an agent I specified—while one of my colleagues did the actual building. So yes, I had the idea and delegated the keyboard workout.

The agent is called “Ask the PMO,” and it helps colleagues across the organization get answers to everyday situations and challenges—basically the friendly PMO buddy you wish you could ping before things get complicated.

In a nutshell, the instructions go something like this:

• Provide expert project management guidance and advice—without making it sound like it came straight from a 400-page methodology handbook.
• Share proven tips, tricks, and best practices that actually survive contact with real project life.
• Help both new and experienced project managers improve their work—because nobody is born knowing how to tame scope creep.
• Answer questions about project processes, methodologies, and frameworks in a way that people can actually use.
• Offer practical solutions to common project management challenges—ideally before the project plan starts quietly crying in the corner.

[Rebekka Heilmann _viadee_]

I have created an Agent that validates if People have uploaded the correct Rovo Fundamentals certificate so we can grant them Rovo Agent Permissions automatically.

The knowledge check is already automated through JSM Forms + Automation. Rovo takes the attachment and checks if the name matches the reporter and the certificate is indeed the correct one (and not a random other course/cert).

This is at least one automated step in a line of Governance tasks, most of which cannot be automated due to lack of API access to Studio/Agents.

[Leonie C_ Breitmoser]

Nice ideas above.

I might adapt some for our environments.I’d like to introduce the most-used agent I’ve created: our cloud onboarding agent. We recently migrated to the cloud and wanted to provide our users with a helpful assistant to guide them through the transition to the new Jira and Confluence interfaces.

Use case: This agent supports end users transitioning from Atlassian Data Center to Cloud by providing user-level guidance only. It helps people use the platform day-to-day without exposing administrative information.

Instructions: You are a helpful Cloud migration assistant designed to support end users transitioning from Atlassian Data Center to Cloud. Your role is to provide user-level guidance only – you help people use the platform, not manage it.

[Tapiwa Samkange]

In my favourite custom agent's own words, em-dashes and all!  "

What I Do

I analyze the portfolio of AI use cases across various teams to identify patterns, optimize resources, and accelerate digital maturity. I process data from our internal tracking system—where functional areas represent teams and specific issues represent AI opportunities—to provide a strategic view of our innovation landscape.

The Problems I Solve

• Fragmentation & Duplication: I identify similar AI needs across different departments (e.g., multiple teams building "Profile Generators") and recommend merging them into shared, scalable solutions to avoid redundant effort.

• Technology Mismatch: I categorize use cases by their tech stack (such as custom assistants, CRM integrations, or built-in AI) and flag when a chosen tool might not be the best fit for the stated problem.

• Strategic Prioritization: I help move initiatives from ideation to implementation by ranking them based on impact and feasibility, ensuring focus remains on high-value "Quick Wins" and "Strategic Bets."

• Governance & Risk: I help monitor data sensitivity and risk levels across the portfolio.

"

[Dave Rosenlund _Trundl_]

Love this!  💙

My "Customer Research & Synthesis" Agent

The Story:
As an Atlassian Community Champion and Global Head of Software & Solutions at Trundl, I spend a lot of time interviewing customers and looking for the "narrative thread" in their success stories. My favorite agent is my Customer Research Agent.

I use this agent to bridge the gap between raw data and a finished story. For example, when I’m preparing for an "Interview with an Expert" article or looking for speaker prospects for Team '26, I have the agent scan through meeting notes, Jira tickets, and Confluence project pages. It doesn't just summarize; it identifies the specific "pain points" and "aha moments" that make for a great Community post.

The Problem it Solves:
It solves the "Information Overload" problem. Instead of me manually re-reading months of project updates to find that one quote or metric that proves ROI, the agent surfaces it for me. It allows me to focus on the storytelling and strategy while the agent handles the discovery and synthesis.

Why it earned a spot in my workflow:
It has turned my research process from a multi-hour deep dive into a 10-minute conversation. It’s like having a research assistant who has read every page in my Confluence instance and remembers exactly where we talked about "Service Management silos" or "AI Governance."