How We Build Value with Atlassian Rovo and Forge
Why Customization is the Key to Real Value from Rovo AI
These days we hear about AI everywhere, all the time. It transforms the way we work, automates tasks, and brings possibilities we hadn’t even thought about a few years ago. Atlassian, with their AI features and especially Rovo, belongs among the leaders in this area. Rovo’s capabilities are huge. Thanks to its deep integration into the Atlassian Ecosystem and its upcoming general availability, it really can make a difference.
However, expectations sometimes don’t match the first experience. The reason is that out-of-the-box tools don’t work instantly for every use case. The key is to understand how Rovo works, what its limits are, and what possibilities exist for extensions. The Rovo Chat with its default agents can already do a lot. But for reliable, value-bringing solutions, customization is needed.
The first step is to create agents in Rovo Studio. There you can define specific instructions, connect various data sources, and use predefined actions to modify data in Atlassian or other apps. This works well, and it is definitely worth trying these options first.
There are use cases where simple prompting and configuration of data sources or actions are not enough. Especially when some features are not yet available, or if you have specific needs or logic where a generic tool is not an option. For that, you can build dedicated Rovo Agents based on the Atlassian Forge platform. Here the options are almost unlimited. This is also the area where we, as Forge and AI experts, bring the most value to our customers.
What we have implemented with Rovo so far:
AQL Helper for Jira Assets
A Marketplace App which turns your natural language questions into precise AQL queries — so you can find, filter, and analyze your assets without writing a single line of code. Whether you’re managing hardware inventory, tracking licenses, or auditing compliance, AQL Helper understands your schema, prepares the correct AQL, and shows you the results in seconds.
Classification of Incoming Marketing Leads
We have various sources of Marketing Leads which we process in Jira and Jira Assets. We built a Rovo agent with instructions, knowledge of our Products and Services, and background information, and incorporated it into Jira Automation. The result is that all incoming leads are automatically assigned to the corresponding Product and Service. This speeds up the marketing process significantly. A similar scenario could also be implemented in Service Desk to classify tickets.
Automatic CRM Contact and Account Assignment
We based our CRM on Jira and Jira Assets. Originally, we created and assigned the Accounts and Contacts for the tickets manually. This was a very time-consuming and error-prone process. Now a dedicated Rovo agent does this for us automatically in the background. For this we needed to create our own Forge-based agent. First, we had to enable working with Jira Assets and all custom fields (which was not available out of the box). Second, we needed to implement special logic for the task. There is a dependency between Contacts and Accounts, and there are various ways to recognize the relevant company from the source information (a ticket). After some fine-tuning, we now have this step fully automated with a very low false prediction rate.
Automatic Signature Removal from Tickets
We use email as a source for Service Desk and Marketing tickets, but incoming messages often include company signatures. These add clutter in Jira descriptions and comments, taking up UI space and interfering with AI tools like Rovo.
Since signatures differ across users and change over time, we needed an AI-driven solution that could recognize them flexibly. Rich Text fields posed another challenge: simple automation broke formatting and images, while handling HTML via API would be too complex.
To solve this, we built a Forge-based Rovo agent that automatically removes signatures while preserving formatting and media.
Let's talk about your use cases and ideas
The pool of possible use cases is almost unlimited. We believe that almost every task in Jira or Confluence done manually today can, at least to some degree, be automated with Rovo. And there are still possibilities we don’t even know yet — ones we will discover later as we adopt the AI-led way of working.
What is your experience? What use cases would you like to automate with Rovo? What worked for you, and what difficulties have you encountered? We would very much like to share this with you, and that is why we organize the Atlassian Rovo in Practice: Interactive Online Workshop. It is not a webinar but a place for discussion and sharing of experiences. During the sessions we have already held, we heard many interesting ideas for utilizing Rovo and also shared our own experience and insights.
If you are curious about Rovo and would like to talk about possibilities for your use case, feel free to register for one of the available timeslots. If none of them works for you, please let us know via comment or message what date and time would fit you, and we can try to arrange that.
Key Contacts:
Petr Sýkora
Matej Štrba
