Data Handling & Retention
What data we collect, how it’s used, how long it’s retained, and how deletion works. This page is written for customer
security and privacy review and avoids proprietary model and taxonomy details.
Last updated: December 16, 2025
Key principles
- Workspace boundary: data access is scoped to the workspace a user belongs to.
- Short-lived message content: message content is processed briefly for analysis and then removed.
- Derived metadata retained: we retain structured results needed for reporting and trend analysis.
- Customer control: admins can connect/disconnect integrations and archive workspaces; deletion requests are supported.
Data categories
Workplace.io handles three main categories of data: account/workspace data, integration data (Slack/Teams), and derived analytics results.
The lists below describe typical data handled by category.
- Account and workspace data: user name, email, role, workspace membership, workspace configuration, and administrative settings.
- Integration data (Slack/Teams): provider identifiers (workspace/team IDs, channel IDs, user IDs), and message events needed for analysis.
- Message content (short-lived): message text is processed during analysis and then removed from storage.
- Derived analytics data: structured outputs such as signal assignments, scores, counts, and aggregated trend data.
- Operational telemetry: service logs and metrics used for reliability and incident investigation (content minimized).
How we use data
We use data to (1) operate the service, (2) connect and maintain authorized integrations, and (3) generate organization-level insights.
We do not publish details of internal model architecture, training data, or proprietary signal taxonomy.
- Operate the service: authentication, workspace membership, billing (if applicable), and support.
- Maintain integrations: connect/disconnect, synchronize events, and handle provider permissions and errors.
- Generate insights: convert short-lived message content into derived metadata and aggregated reporting.
- Security operations: monitor availability, investigate incidents, and respond to abuse.
Retention overview
Retention depends on data type. Message content is intentionally short-lived; derived results and operational records may be retained longer
to support reporting, auditability, and service integrity. If your organization requires a specific retention schedule, contact us.
Short-lived message content
Message content is processed for a short period during analysis and then removed from storage. We retain primarily derived metadata and
aggregated results used for dashboards and trends.
- Message content: short-lived during processing; removed after analysis completes.
- Derived metadata: retained to support dashboards, trends, and product functionality.
- Account/workspace data: retained while the account/workspace is active; updated as membership changes.
- Operational logs: retained for troubleshooting, security investigation, and reliability; sensitive content minimized.
Deletion and customer requests
Customers may request deletion of workspace data. Deletion requests are scoped, verified, and executed according to an internal process to
prevent accidental loss and ensure correct tenancy boundaries.
- Disconnect integrations: stops future ingestion and processing for that provider connection.
- Archive workspace: disables access and removes the workspace from selection; does not delete underlying data by itself.
- Delete workspace data: available upon request; we coordinate scope and timing with an authorized workspace admin.
- Export support: upon request, we can help provide data needed for audits or transition planning.
Data access and least privilege
Access is scoped by workspace membership and roles. Administrative functions (integrations, billing, user management, archiving) are limited
to authorized roles. Internal access to production systems and sensitive data is restricted.
Where data is processed
Workplace.io uses DigitalOcean for application hosting and managed databases. Classification and culture inference workloads run on AWS using
Amazon SageMaker. See Subprocessors for a vendor directory.