Atlassian and OpenAI expand partnership to bring frontier models into Jira and Confluence

Atlassian and OpenAI expand partnership to bring frontier models into Jira and Confluence

The deal connects OpenAI's frontier models to Atlassian's Teamwork Graph, betting that AI agents work better when they know how your company actually operates

Atlassian and OpenAI are deepening a partnership that started in 2023. The expanded deal, announced October 6, 2026, brings OpenAI’s GPT-6 family of models into Atlassian tools like Jira and Confluence.

The goal is simple to state and hard to pull off: give AI agents enough company context to be useful, not just articulate.

What the expanded deal includes

The centerpiece is Rovo, Atlassian’s AI tool. Rovo pairs OpenAI’s models with something Atlassian calls the Teamwork Graph.

Think of the Teamwork Graph as a map of how work flows inside an organization. It links teams, projects, documents, and decisions into a single contextual layer.

According to the announcement, the combination is meant to support what the companies describe as organizational reasoning. That means AI that can connect a Jira ticket to the Confluence page explaining why the ticket exists in the first place.

The partnership also introduces Atlassian MCP Server connectors. These are designed to let data move between ChatGPT, Codex, and Atlassian’s tools.

Advertisement

The practical upshot: a developer working in Codex could, in theory, pull relevant project context from Jira without switching tabs.

According to the research findings on the announcement, OpenAI’s models are intended to let users tap AI capabilities matched to specific project needs, rather than a one-size-fits-all assistant.

Atlassian is already eating its own cooking

Atlassian is not just reselling OpenAI’s technology. It is using it heavily in-house.

More than 3,000 developers at Atlassian use Codex, OpenAI’s coding tool, for software tasks.

Atlassian has also expanded its internal use of ChatGPT Enterprise.

How we got here

The relationship between the two companies dates back to 2023. Since then, the enterprise AI conversation has shifted from novelty chatbots toward agents that can take action inside business software.

Atlassian’s pitch is that it already holds those documents. Jira tracks the work, Confluence holds the knowledge, and the Teamwork Graph connects the two.

OpenAI, for its part, supplies the frontier models. The expanded deal effectively marries one company’s data layer with the other’s reasoning engine.

What this means for enterprises and the AI market

For joint customers, the most immediate change is that AI features are expected to show up directly inside the workflows they already use.

The MCP Server connectors will be a key test. Seamless data sharing is a feature right up until it shares something it should not.

The internal Codex numbers at Atlassian offer a benchmark. If more than 3,000 of the company’s own developers find the tooling useful, that is a reasonable proxy for how software-heavy customers might respond, and a useful figure to compare against future adoption disclosures.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.
Atlassian and OpenAI expand partnership to bring frontier models into Jira and Confluence
Atlassian and OpenAI expand partnership to bring frontier models into Jira and Confluence

The deal connects OpenAI's frontier models to Atlassian's Teamwork Graph, betting that AI agents work better when they know how your company actually operates

Atlassian and OpenAI are deepening a partnership that started in 2023. The expanded deal, announced October 6, 2026, brings OpenAI’s GPT-6 family of models into Atlassian tools like Jira and Confluence.

The goal is simple to state and hard to pull off: give AI agents enough company context to be useful, not just articulate.

What the expanded deal includes

The centerpiece is Rovo, Atlassian’s AI tool. Rovo pairs OpenAI’s models with something Atlassian calls the Teamwork Graph.

Think of the Teamwork Graph as a map of how work flows inside an organization. It links teams, projects, documents, and decisions into a single contextual layer.

According to the announcement, the combination is meant to support what the companies describe as organizational reasoning. That means AI that can connect a Jira ticket to the Confluence page explaining why the ticket exists in the first place.

The partnership also introduces Atlassian MCP Server connectors. These are designed to let data move between ChatGPT, Codex, and Atlassian’s tools.

Advertisement

The practical upshot: a developer working in Codex could, in theory, pull relevant project context from Jira without switching tabs.

According to the research findings on the announcement, OpenAI’s models are intended to let users tap AI capabilities matched to specific project needs, rather than a one-size-fits-all assistant.

Atlassian is already eating its own cooking

Atlassian is not just reselling OpenAI’s technology. It is using it heavily in-house.

More than 3,000 developers at Atlassian use Codex, OpenAI’s coding tool, for software tasks.

Atlassian has also expanded its internal use of ChatGPT Enterprise.

How we got here

The relationship between the two companies dates back to 2023. Since then, the enterprise AI conversation has shifted from novelty chatbots toward agents that can take action inside business software.

Atlassian’s pitch is that it already holds those documents. Jira tracks the work, Confluence holds the knowledge, and the Teamwork Graph connects the two.

OpenAI, for its part, supplies the frontier models. The expanded deal effectively marries one company’s data layer with the other’s reasoning engine.

What this means for enterprises and the AI market

For joint customers, the most immediate change is that AI features are expected to show up directly inside the workflows they already use.

The MCP Server connectors will be a key test. Seamless data sharing is a feature right up until it shares something it should not.

The internal Codex numbers at Atlassian offer a benchmark. If more than 3,000 of the company’s own developers find the tooling useful, that is a reasonable proxy for how software-heavy customers might respond, and a useful figure to compare against future adoption disclosures.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.