Atlassian this week revealed it has made generally available Rovo Dev, an artificial intelligence (AI) agent that can be used to not only write and review code but also plan projects.
Announced at the Atlassian Team ’25 Europe conference, Atlassian also announced it is bundling Rovo Dev with Bitbucket continuous integration/continuous deployment (CI/CD) platform, the Compass catalog for managing software components and the DX engineering metrics platform it acquired last month into a single offering dubbed Atlassian Software Collection.
Rovo Dev is embedded within a command line interface (CLI) and can also be used to review code residing in Bitbucket or GitHub based on acceptance criteria defined by a DevOps team. There will also soon be a deeper integration with Jira to enable coding tasks to be assigned to Rovo Dev to complete. Based on the same AI framework and Teamwork Graph technologies used across the entire Atlassian portfolio, Rovo Dev also makes it simpler to connect an AI tool for developers with Bitbucket, Compass and Jira project management software running on a cloud service hosted by Atlassian.
Aidan Cunniffe, principal product manager for Software Collection at Atlassian, said the goal is to build a world where human engineers are able to be a lot more productive than they were before, and have a better time coding by delegating tedious tasks to an AI agent. It will also become a lot simpler to integrate disparate software development tools using AI agents that are communicating with each other via the Model Context Protocol (MCP) developed by Anthropic, he added.
It’s not quite clear how many agents software engineers might need to ultimately deploy, but the next major challenge will clearly involve orchestration. Each developer is likely to have their own personal agents to automate tasks, in addition to, for example, being able to invoke an AI agent that has been assigned a specific task to automate on behalf of the entire software development team.
Collectively, AI agents will not only accelerate the pace of new software development but also make it a lot easier to modernize legacy application code, said Cunniffe.
Each DevOps team will need to determine for itself to what degree to rely on AI agents to automate tasks, but hopefully, software development is about to become far less tedious. Software engineers will, of course, still need to review any code generated using AI tools before promoting it into a production environment.
Ultimately, it’s not so much a question of whether code will be created using AI tools as it is determining how best to take advantage of them. Regardless of approach, the one thing that is certain is that the amount of software that can be developed without increasing the overall headcount of a software engineering team is about to substantially increase. The challenge now is finding a way to manage all the code as it moves through DevOps pipelines from a previous era of software development that was not designed to handle that level of volume.
