Atlassian and OpenAI deepen GPT-6 integration with Rovo and Jira workflows
Atlassian and OpenAI expand their partnership to connect GPT-6 models with Rovo and enterprise project context, while deeper Jira agent workflows remain under development.
The expanded agreement connects OpenAI models with Atlassian’s Teamwork Graph. Existing context, supported-action and Jira-to-Codex integrations differ from deeper autonomous execution still in development.
Atlassian and OpenAI announced an expanded partnership on 6 October 2026, bringing GPT-6 integration with Rovo and deeper connections to Jira-based workflows. The agreement gives Atlassian broader access to OpenAI models, including GPT-6 Astra, while connecting AI assistants to enterprise information held across Jira, Confluence and other Atlassian products.
In their official announcement, the companies said OpenAI models would support agents across Atlassian’s platform. Rovo combines those models with the Teamwork Graph, which connects people, projects, documents and decisions. The expanded arrangement builds on a partnership established in 2023.
Rovo connects model reasoning to project context
The Teamwork Graph supplies organisational context that a general-purpose model would not otherwise possess. For example, a product manager assessing launch readiness could ask Rovo to examine Jira work items, relevant Confluence documents and project discussions to identify blockers or missed milestones.
OpenAI says Atlassian can incorporate newer reasoning capabilities through its APIs as models become available. Atlassian’s partnership explanation also confirms that OpenAI models operate alongside other models across the platform. The agreement does not make Rovo exclusively dependent on OpenAI.
ChatGPT and Codex can use Atlassian context
The partnership extends to ChatGPT and Codex through Atlassian integrations that provide access to authorised work information. Atlassian’s plugin and Model Context Protocol connections allow users to retrieve Jira work items, Confluence content and development documentation without manually copying information between systems.
Atlassian says more than 3,000 of its developers already use Codex across terminals, development environments and code-review workflows. That is an internal adoption figure reported by the company, not an independently verified measure of productivity.
The company also describes a pinned Atlassian Home experience that surfaces assigned work, recent Loom recordings, projects and Bitbucket pull requests within supported OpenAI workflows.
Deeper Jira agent execution remains planned
The companies are exploring capabilities that would let AI agents take assigned Jira work, execute development tasks, record progress and return results for human review. Atlassian has also discussed linking those activities with its DX engineering analytics platform to assess development performance.
These planned capabilities go beyond the integrations available today. Atlassian’s current integration guide documents permission-aware context access, supported actions for review and Jira-to-Codex handoffs that prepare implementation plans. Autonomous work-item pickup, test execution and session synchronisation remain future-facing, without a confirmed general-availability date.