A Feedback Dataset is the top-level container for feedback inside an organization. Every Feedback Record belongs to exactly one dataset, and every source writes into a single dataset.
When to create a new dataset
Create one dataset per logically separate group of feedback. Common patterns:
- By product line (e.g. “Mobile devices”, “Web apps”)
- By stakeholder group (e.g. “Customers”, “Employees”)
- By region (e.g. “Europe”, “North America”)
Workspace access
Datasets live at the organization level but are exposed to workspaces through an access list. Each workspace can only access one dataset.
Manage dataset access from Settings → Organization → Datasets:
- Create new datasets
- Rename or archive datasets
- Add or remove workspace access
- Delete all records in a dataset
Only Owners and Managers can manage datasets. Workspace members see the datasets their workspace has access to under Analyze.
Archiving
Archiving a dataset hides it from default views but does not delete its records. Use it for one-off programs that have ended.
Delete all records
Delete all records empties a dataset while keeping the dataset itself. Open Settings → Organization → Datasets, click Manage on the dataset, then click Delete all records.
This cannot be undone, and it is not limited to your workspace. It permanently deletes every feedback record in the dataset — including its AI enrichment and the topics generated from it — for every workspace the dataset is shared with.
The dataset and its sources are kept, so new feedback keeps arriving. Topics are generated again once new feedback comes in, and charts built on the dataset are empty until then.
To confirm, type the dataset’s exact name. Only Owners and Managers can do this; everyone else sees the button disabled.
Deletion runs in the background, so records may still show up for a few minutes after you confirm.
This is what separates it from archiving: archiving keeps the records and hides the dataset, deleting keeps the dataset and removes the records.