OpenAI and Atlassian Bring Jira Work Into ChatGPT

OpenAI and Atlassian announced an expanded partnership on October 6 (local time). At its core is an effort to connect AI with the work information accumulated inside companies. Under the new agreement, OpenAI models including the GPT-6 family will power the Atlassian platform and Rovo, an AI agent that handles tasks on its own. The goal is to support teams from planning through development and task completion.

The two companies first began working together in 2023. The new agreement gives Atlassian broader access to OpenAI models, including GPT-6 Astra and the GPT-5.6 family. Rovo can also take on the new reasoning capabilities of upcoming OpenAI models — the ability to work through a problem and arrive at an answer on its own — via APIs, the mechanism developers use to plug a service into their own products.

What work information gets connected to the AI

Rovo pairs OpenAI models with Atlassian’s Teamwork Graph. Teamwork Graph is a feature that maps the relationships among people, projects, documents, and decisions. It passes along to the AI not just individual pieces of work information, but how those pieces connect to one another.

ChatGPT and Codex can also pull in work information from Atlassian. To make that possible, plugins have been built that retrieve data from Atlassian and Teamwork Graph. Their job is to connect project information, documents, and development context from the workflows people already use.

Specifically, they pull in Jira work items, Confluence documents, and information about people. That information is included in the prompt, the instructions given to the AI. When users assign a task to ChatGPT or Codex, they can hand over internal work context along with it. When the plugins retrieve information, each person’s existing permissions still apply.

Will AI work be managed in Jira?

Customers who use both companies’ products can use this integration. Atlassian said more than 3,000 of its own developers are already using Codex. According to the company, they use Codex throughout the development process — in the terminal, where commands are typed in, inside the programs they use to develop software, and in code review.

The two companies are also exploring a deeper integration that would manage AI work inside Jira. The idea is to hand a task to an AI and then track its progress and results in Jira alongside everything else. Beyond the information-linking features announced now, they are looking at whether to extend the partnership to assigning and tracking the work the AI takes on.

What remains to be seen from here is whether the Jira integration under review takes shape and, if so, what it looks like. The key question is whether a concrete way emerges to handle the progress and results of tasks assigned to AI within a team’s own task management.