Atlassian has confirmed a major change to how customer data may be used across its cloud products, announcing that it will begin collecting customer metadata and selected in-app content by default to improve its expanding AI portfolio from August 2026.
The decision marks a significant shift in positioning for the collaboration software giant and reflects a wider transformation underway across enterprise software as vendors compete to strengthen generative and agentic AI capabilities.
For the roughly 300,000 organisations running Atlassian Cloud, the update introduces a new reality in which the balance between innovation, privacy, and pricing becomes more explicit than ever.
In a statement provided to UC Today, an Atlassian spokesperson said the proposed changes will “improve Atlassian apps and AI experiences”.
“We’re invested in delivering the best AI experiences for our customers – and with richer, more diverse customer data and usage patterns, we can deliver enhanced AI capabilities that unlock greater value across an organisation.
“Our customers trust us with their most business‑critical work, and we take that responsibility seriously. Alongside this change, we’re strengthening our safeguards by de‑identifying and aggregating the data used for these improvements, and we’re giving organisations more control through new in‑app data contribution settings. This includes the ability for all customers to fully opt out of in‑app data contribution.”
A Sign of the Times?
The new policy introduces a tiered model that ties data contribution directly to subscription levels.
From the company’s perspective, the change is designed to accelerate the performance of AI features embedded across tools such as Jira and Confluence, particularly its Rovo and Rovo Dev assistants.
This shift mirrors a broader trend across enterprise technology.
As AI capabilities move from experimental add-ons to core functionality, vendors increasingly view real workplace usage signals as essential to improving accuracy, relevance, and automation.
Atlassian’s decision reflects the growing reality that enterprise software is becoming both the platform for productivity and a key source of signals used to enhance AI systems.
What Data May Be Collected
Atlassian’s new approach separates data into two categories: metadata and in-app content, each with different default settings.
Metadata refers to de-identified signals about how teams work rather than the content they create.
These signals include metrics tied to project management, service delivery, and workflow complexity.
The company has confirmed that metadata collection will be mandatory for customers on Free, Standard, and Premium plans, with no opt-out available on those tiers.
Alongside this, Atlassian may collect in-app content, including user-generated material such as Jira issue descriptions and Confluence pages, to improve AI features.
The default settings for this category vary depending on subscription level.
For Free and Standard customers, in-app data collection will be enabled by default but can be disabled by administrators.
Premium customers will see the setting turned off by default, while Enterprise customers retain full opt-out control.
The tiered approach makes clear that the level of governance organisations can exercise over their data is closely tied to subscription level and configuration.
The company has also confirmed that contributed data may be retained for up to seven years. Atlassian says in-app data will be removed within 30 days following deletion or opt-out, and any models trained on that data will be retrained within 90 days to remove the contribution.
While these safeguards are intended to reassure customers, long retention windows are likely to be carefully reviewed by organisations with strict governance and audit requirements.
Not all customers will be affected.
Organisations using customer-managed encryption keys, Government Cloud, Isolated Cloud, or operating under HIPAA obligations are excluded from the data contribution programme.
These environments are typically associated with specialised or higher-cost deployments, reinforcing the link between privacy and pricing that runs throughout the policy.
The New Privacy Versus Price Equation
At the centre of the announcement is a new commercial reality for enterprise software buyers.
Smaller organisations using lower-cost plans will face limited ability to opt out of certain data contributions, while larger enterprises paying for premium tiers retain greater control.
This emerging model reflects the economic pressures of the AI race.
Developing advanced AI features requires large volumes of real-world usage signals, while customers increasingly expect intelligent automation and generative capabilities to be embedded into the tools they already use.
The result is a shift in the pricing logic of enterprise software.
Organisations are effectively being asked to choose between contributing data to help improve AI features or investing in higher-tier deployments that provide stronger governance controls.
For many IT and procurement leaders, the cost of data control is becoming a key consideration in the AI era.
Scrutiny, Preparation, and What Comes Next
Despite Atlassian’s assurances that contributed data will be de-identified and aggregated, the policy is likely to be closely examined by customers and regulators as part of broader discussions around AI governance and transparency.
Even without direct identifiers, operational data can reveal patterns about organisational workflows and delivery processes, which many businesses consider commercially sensitive.
With the rollout scheduled for August 2026, organisations now have time to evaluate the implications. review subscription tiers, update governance documentation, and assess whether higher-tier deployments are required to meet internal compliance requirements.
The decisions made may ultimately become part of a wider conversation about the long-term cost of data control as AI becomes embedded in everyday workplace software.
