English 
On April 8, Atlassian made Rovo Dev generally available. This is a context aware AI agent built to support software development by using data that already exists inside tools like Jira.
The idea is not just AI helping with code. It is AI understanding the full context around the work. Requirements. tickets. past changes. documentation. all of it.
Rovo Dev connects across Atlassian’s ecosystem. That includes Confluence, Bitbucket, and Jira Service Management.
All of this data sits inside something Atlassian calls the Teamwork Graph. That is basically the shared data layer across their tools. Rovo Dev uses that to understand what is going on before it starts doing anything.
From Context to Execution
Earlier, Rovo Dev was only available through a command line interface. Now it runs directly inside Jira.
That changes how people use it.
A developer can start from a Jira work item. The AI reads the requirement. It looks at history. It understands related tasks.
Then it generates an execution plan.
After user approval, it can actually move forward. It makes code changes. Runs tests. Then creates a pull request linked back to the Jira ticket.
So instead of jumping between tools, a lot of the flow stays inside Jira.
こちらもお読みください： オープンAIが過去最高の$1,220億円の資金を獲得、世界のAI事情を再構築へ
Cutting Down Tool Switching
Traditional workflows are fragmented.
Planning happens in Jira. Coding happens in an IDE. Repositories sit in Bitbucket or GitHub. Pull requests happen elsewhere.
Developers keep switching context. That costs time.
With Rovo Dev, Atlassian is trying to collapse that flow.
You start in Jira. You stay in Jira for a large part of the process. The AI handles the transitions in the background.
Developers can spend more time on design decisions and reviews instead of repetitive steps.
Visibility for Non Developers
Another piece here is visibility.
When Rovo Dev creates plans and progresses through tasks, everything is linked back to Jira work items.
That means product managers and designers can see what is happening without digging into code repositories.
Progress is easier to track. Context is shared across roles.
There is also automation on top of this. Jira rules can trigger Rovo Dev automatically when certain types of work items are created.
So in some cases, the process starts without manual input.
How It Handles Testing and Safety
Testing runs in a cloud based sandbox.
The system clones the repository into that environment. Runs tests there. Keeps it isolated.
Permissions are tied to the user who starts the session. So the AI does not get unlimited access. It operates within defined boundaries.
Pull requests still go through human review before merging. That part has not changed.
So it is not full automation. It is assisted execution with checkpoints.
Pricing and Usage
Rovo Dev is priced at 2,730 yen per developer per month.
Each user gets 2,000 Rovo Dev credits monthly as part of that plan.
What This Means in Practice
This is another step toward AI sitting inside workflows, not outside them.
Instead of opening a separate AI tool, the AI is embedded where work already happens.
The key difference here is context.
Because Rovo Dev pulls from Jira, Confluence, and code history, it is not guessing. It is working with actual project data.
That reduces friction. It also reduces the gap between planning and execution.
Bigger Shift Behind This
Tools are starting to behave less like tools and more like participants in the workflow.
Rovo Dev does not just assist. It plans. executes. tests. creates pull requests.
But still within guardrails.
For teams already using Atlassian’s stack, this kind of integration can change how work flows day to day.
Less switching. Less manual setup. More focus on higher level thinking.
That is where most of these AI dev tools are heading.
Not replacing developers.
Reducing everything around them that slows them down.
日本語