Atlassian and OpenAI expanded their partnership on 6 October to put GPT-6 family models into agents across Atlassian’s platform and Rovo, extending a collaboration that began in 2023, OpenAI said.
Key points
- The agreement gives Atlassian expanded access to GPT-6 Astra and the GPT-5.6 series, OpenAI says.
- Rovo combines OpenAI models with Atlassian’s Teamwork Graph, which connects people, projects, documents and decisions.
- More than 3,000 Atlassian developers use Codex in terminals, IDEs and code review workflows, according to OpenAI.
- The companies are exploring deeper Jira integrations for assigning work to AI agents and reviewing their results.
GPT-6 Astra joins Atlassian’s model access
Under the new agreement, OpenAI models will power agents in Atlassian products, including Rovo. Atlassian also gains wider access to GPT-6 Astra and the GPT-5.6 series, according to OpenAI’s announcement. The arrangement places those models inside software used to organise work, while Rovo supplies context from Atlassian’s products.
The companies began working together in 2023. The expanded agreement covers both the models Atlassian can use in its own platform and connections that let customers bring Atlassian information into OpenAI products. OpenAI says it, in turn, continues to use Jira to manage critical workflows across the company.
Jamil Valliani, Atlassian’s Head of Product, AI, described the purpose of the expansion as grounding OpenAI models in customers’ day-to-day work through the Teamwork Graph. Nikunj Handa, OpenAI’s Head of Product, API, said the companies want teams to use that context to take action on their work.
Rovo draws context from the Teamwork Graph
Atlassian’s Teamwork Graph connects people, projects, documents and decisions. Rovo pairs that information with OpenAI models, which Atlassian can access through OpenAI APIs as the models advance, the announcement said. The proposed use is specific to work already recorded in company systems, rather than a general-purpose exchange with a chatbot.
OpenAI offered a product-launch scenario as an example. A manager could ask Rovo whether a team is on schedule. Rovo could draw on Jira tickets, Confluence documents and discussions to find engineering obstacles, missed milestones and decisions awaiting attention. OpenAI models would then help produce an assessment of launch readiness and recommend next steps. The scenario describes a possible workflow, rather than a reported result from a customer deployment.
The companies are also exploring closer Jira integrations. Those would let teams assign tasks to AI agents, follow their progress, record decisions and review the resulting work, according to OpenAI. Atlassian’s DX platform measures developer productivity and engineering performance. OpenAI says pairing it with the proposed capabilities could help engineering leaders assess development speed, cycle time and developers’ experience.
More than 3,000 Atlassian developers use Codex
Atlassian’s own use gives the partnership an existing deployment alongside the planned product integrations. More than 3,000 of its developers use Codex across terminals, integrated development environments and code review workflows, OpenAI said. Atlassian has also been broadening its use of ChatGPT Enterprise.
Atlassian plugins drawing on the Teamwork Graph can give Codex users access to work items and technical documentation. OpenAI says developers can use that material while writing, testing and shipping software. The collaboration also reaches customers using ChatGPT and Codex outside Atlassian’s own products: Atlassian and Teamwork Graph command-line plugins can connect those tools to project information, documentation and development context, subject to permissions.
Atlassian has launched a plugin extension that places Jira work items, Confluence content and people in ChatGPT and Codex prompts. Its pinned Atlassian Home has also surfaced assigned work, recent Looms, projects and Bitbucket pull requests.
