Atlassian’s recent policy change to collect in-app data for AI training amid security concerns and limited user opt-out options raises questions over customer trust and its impact on the company’s upcoming earnings report.
Atlassian’s move to begin collecting metadata and in-app data from Jira, Confluence and other cloud products for AI training has landed at an awkward point for the company. The policy took effect on 16 August and gives enterprise customers an opt-out, but users on Free, Standard and Premium plans do not get that choice. That matters because the software group serves about 300,000 customers globally, and trust over data handling is central to its enterprise pitch.
The timing is sensitive because Atlassian is leaning harder into AI across its product set. According to its product pages, Confluence AI is built around Rovo to connect knowledge across Confluence, Jira, Slack and other workplace applications, while Jira AI uses the same assistant for search, automation and content generation. Atlassian describes Rovo as an AI assistant that helps teams find, learn and act on information across its own tools and third-party apps.
Security concerns have added to the scrutiny. In early August, two security firms showed that Rovo could be manipulated into collecting a logged-in user’s Jira or Confluence data and sending it to an external server. The company has confirmed that only one of the two attack routes has been fixed, leaving a question mark over how much control customers can realistically expect while the new data policy is in force.
Even so, the company continues to frame AI as a productivity layer rather than a standalone model business. Atlassian has recently expanded Confluence with partner agents and broader Rovo Studio workflows, which it says can help teams generate product requirements, status updates and onboarding material from existing knowledge. It has also said its software is seeing more than five million agent calls a month, while a Jira summer release due to be fully deployed by the end of August will add a Delivery Agent for more proactive automation.
That strategy sits alongside a wider push towards cloud software and away from legacy data-centre products. Atlassian has guided for fiscal 2027 subscription annual recurring revenue growth of about 18 per cent and cloud revenue growth of around 25 per cent, while its data-centre business is expected to contract by roughly 17 per cent. The company is also positioning itself as the orchestration layer for tools such as Claude Code, Cursor and GitHub Copilot, rather than trying to outbuild the major AI model providers.
For investors, the tension is between product momentum and customer confidence. The shares had already surged after a strong earnings report on 6 August, and the recent pullback comes after a steep move higher. Insider transactions have pointed in different directions too: director Anutthara Bharadwaj sold about 8,000 shares on 14 August, while co-founder and chief executive Mike Cannon-Brookes announced a stock purchase plan of up to $250 million on 6 August. The next key test comes on 29 October, when Atlassian reports again and investors can judge whether the AI push is translating into durable demand, or whether the data policy has started to weigh on sentiment.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
