Forums

Articles
Create
cancel
Showing results for 
Search instead for 
Did you mean: 

How the Atlassian product management team uses Rovo Agents

Have AI-powered agents changed how you work?

At Atlassian, almost all of our 450+ product managers now use AI tools weekly, saving nearly 40 minutes per day on average. Part of this transformation is happening with agents, our new virtual team mates.

Here are four agents that have changed different aspects of our product discovery process:

1️⃣ 🔀 Decision Director: Enhances decision-making by reviewing documents, making suggestions, connecting teams with subject matter experts, helping improve decision clarity, and ensuring alignment with whatever decision framework the team is following.

2️⃣ 🗒️ Product Requirements: Assists in drafting and reviewing product requirement documents, aligning teams with common templates, using specific sources as context and focusing on problem articulation.

3️⃣ 🔍 Customer 360: Before a customer call, this agent pulls together everything from Jira, Confluence, Salesforce, Slack, and internal systems—giving PMs a complete view of a customer in seconds.

4️⃣ 📢 Jira Theme Analyser: We get thousands of feedback items weekly. This agent automatically scans Jira work items to find themes, posting daily summaries to Slack or publishing reports in Confluence. It's helping teams spot top requests, support issues, and trends—without drowning in the data.

👉🏽  What to lean more? I've gone into detail with examples of each agent with screenshots/demos here: https://sherifmansour.medium.com/four-agents-atlassians-product-managers-use-to-improve-their-product-discovery-workflows-2201b99b68af 
Have you had a go at using Rovo agents? Any ideas to share for how it's changed how you work? What can we do to make them better? Love to hear our thoughts 

6 comments

Brandon Davies
Community Champion
July 3, 2025

Thank you @Sherif Mansour , always love to see your articles and learnings, is there any chance we could see some of the code from the "Customer 360" agent as I want to learn and understand what exactly makes it different to using the out of the box custom Rovo Agent through text, is it just the extra connectors through API's or are there more functions coded? My experience with the custom rovo creator NLP version has been very hit or miss. Thank you!

Like # people like this
Brita Moorus
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
July 3, 2025

Customer 360 is a great idea - this can really come in handy before customer calls. Thank you for sharing this! ✨😊

Like # people like this
Sherif Mansour
Atlassian Team
Atlassian Team members are employees working across the company in a wide variety of roles.
July 4, 2025

@Brandon Davies I'll check specifically on that agent - was written by our internal IT team, but in general we do have examples of code agents e.g. Refapp: https://bitbucket.org/atlassian/rovo-weather/src/main/ also here are sample apps: https://developer.atlassian.com/platform/forge/example-apps-rovo/ 

 

But a high-level things you can do in a code-based agent that you can't do with a no-code agent: 

1. Connect to internal systems , databases - whatever you can do with code.

2. Add custom code logic to how your agent runs or where it appears (e.g. you could also combine the agent with other Confluence/Jira forge modules to expose functions of it in other places)

3. You need access to call external APIs, webhooks etc.

4. Add some very deterministic logic to how the agent works

There are lots of examples of Marketplace vendors building agents or another consideration could be to just expose an "action" (a specific task an agent can do) so that anyone with no-code can re-use that action in their agent. See: Customer created actions. E.g. Our customer 360 agent also exposes an action to "get customer domains" , so for any customer the agent can return all the domains they own - which is an internal system call. 

Finally, "My experience with the custom rovo creator NLP version has been very hit or miss" - oo would love to hear more? Any feedback would be great so we can improve it. 

Like # people like this
Brandon Davies
Community Champion
July 4, 2025

@Sherif Mansour Thank you very much Sherif! Regarding feedback, I would have to write up a well documented and accurate reenactment to be useful feedback instead of rambling :D

Is there somewhere I could send this feedback to? 

I could get working on that soon.

Like # people like this
Dr Valeri Colon _Connect Centric_
Community Champion
July 15, 2025

Even when @Sherif Mansour posts are short they're packed with knowledge. Stealing these links for our resource library.

Like Amanda Barber likes this
Dr Valeri Colon _Connect Centric_
Community Champion
July 15, 2025

What?! Sherif is on Medium Magazine. Smashing the follow button.

Like Amanda Barber likes this

Comment

Log in or Sign up to comment
TAGS
AUG Leaders

Atlassian Community Events