Canonical assistant reference · updated August 8, 2026
Aria, the AI assistant built into Stelaah
Aria is Stelaah's in-product AI assistant. It is designed to help people work with permitted workspace context while keeping consequential decisions and generated output subject to human review.
Documented capabilities
Depending on the workspace, plan, and current product availability, Aria can assist with drafting, summarizing records, answering questions from available workspace context, creating meeting notes, and helping with supported actions. The Aria product reference is the detailed feature source.
How context works
Useful assistance depends on the information a user is authorized to access. Workspace context can include related records such as clients, projects, tasks, documents, and meetings when the product makes those records available to the request.
Permissions
Aria does not create a separate entitlement to workspace data. Requests and assisted actions must remain within the user's access and the workspace's authorization model. Users should not ask Aria to disclose information they are not permitted to access.
Privacy and data handling
AI data handling is governed by Stelaah's current privacy policy, AI terms, and applicable agreements. Do not infer model-training, retention, residency, or certification commitments that are not stated in those sources.
Automation and actions
Aria may assist with supported actions, but an AI suggestion is not evidence that an action was completed. Important changes should expose their target, permission boundary, and outcome and should be reviewed by the responsible user.
Summaries and generated content
Summaries and drafts can omit context or contain errors. Review source records before relying on generated output or sending it to a client. Legal, financial, employment, privacy, and safety decisions require appropriate professional and human review.
Predictions and background intelligence
No prediction, score, or background recommendation should be assumed unless the current product explicitly presents it and identifies its basis. Aria should not be described as making autonomous business decisions in the background.
Safety boundaries
- AI output is assistance, not guaranteed fact.
- Workspace permissions still apply.
- Users remain responsible for consequential decisions and external communication.
- Availability varies by feature, plan, configuration, and release.
- Undocumented capabilities should not be inferred.