The administrative control for AI features in Rally has moved to the subscription level, giving organizations a faster and more centralized way to switch capabilities on or off. Instead of filing a Support ticket, a subscription administrator can enable or restrict AI tools across the subscription. Note that Rally Anywhere does not include these AI features, and they cannot be activated inside a Sandbox environment. This article explains the prerequisites, default behaviors, and a clear inventory of which AI features do and do not use customer data.
Centralizing AI configuration at the subscription tier means settings are not adjusted at the workspace or project level; however, once enabled for a subscription, administrators can still control access at the individual feature and user level. Administrators should understand the approval steps that are required before turning capabilities on, and how default access works depending on whether the AI Module was previously enabled through Support.
Prerequisites and administrative responsibilities
To enable the suite of AI capabilities, the subscription administrator must perform the changes at the subscription level and accept the Rally AI Terms and Conditions. Because AI controls are centralized, you cannot configure them per workspace or project; instead, permissions are scoped to specific features and to users. If your organization never used a Support ticket to enable the AI Module, the ability to ask questions via the Rally AI Chat Bot is switched on by default. Conversely, if your subscription previously worked with Support to enable the module, your subscription will retain the AI features that were available at that earlier activation point.
Default and customizable chat bot behavior
There are two important distinctions to know about the Rally AI chat capabilities. First, the basic chat bot that answers questions from Rally documentation is available without exposing customer data. Second, subscription administrators can opt in to extended chat capabilities that make use of customer artifacts and attachments to provide richer, contextual responses. The sub administrator can also fine-tune which users or teams can access these enhanced capabilities, enabling a granular control model after the subscription-level switch is made.
Environment and scope restrictions
Keep in mind two environment restrictions: Rally Anywhere is not supported for these AI features, and you cannot enable them inside a Sandbox environment. These limits protect experimental or remote deployments and ensure that production data handling policies remain consistent for supported platforms.
AI features that use customer data
The following Rally features can leverage customer data to generate responses, summaries, or content. All listed items use the Gemini 3.0 Flash model, which is proprietary, and none of these features store customer inputs for model training. Key items include: Advanced Add New with AI Assistant to decompose features into user stories; the AI Writing Assistant in the Detail Editor to accelerate content creation and refinement; AI Attachments as Context that allow documents to be used as context; the Team Health Widget for AI-generated summaries based on flow metrics; Rally AI Chat Bot – Create User Stories which prompts for required information; the Query Assistant that produces queries for custom views; the Throughput/Velocity and WIP Chart Interpreter which analyzes trends and provides observations and recommendations; the AI-Powered Train Health Widget that summarizes release train project health; and the AI-Driven Artifact Breakdown into Children that creates child artifacts and can use attachments as context.
Data handling assurances
For all of the features listed above, Rally documents that customer data is not used to train the underlying model. The combination of a proprietary language model and explicit non-training data policies aims to balance functional value with privacy safeguards. Administrators should still review their internal compliance policies before enabling features that reference customer artifacts or attachments as context.
AI features that do not use customer data
Separately, some AI capabilities provide assistance without accessing or storing customer-specific information. The most notable example is the Rally AI Chat Bot – Basic Questions, which answers questions using Rally public documentation as the knowledge source. This capability also runs on Gemini 3.0 Flash, is proprietary, and does not store customer inputs for model training. It is a lower-risk option for teams that need quick guidance about how Rally works without exposing internal data.
In summary, subscription administrators now have direct control to turn AI features on or off, accept the necessary terms, and choose whether to enable contextual, data-aware tools for users. Understand the environment exclusions (Rally Anywhere and Sandbox), the default behaviors for chat capabilities, and the clear separation between AI features that use customer data and those that do not. Proper governance and informed enablement will help teams take advantage of Rally’s AI tools while maintaining their data policies.
