This document explains what data we collect, how authentication works, and what each feature accesses.
We use Clerk for authentication. When you sign in, Clerk sets strictly necessary session cookies (__clerk_*, __session). These are required for the platform to function and are not subject to cookie consent. No passwords are stored on our servers. Survey respondents do not sign in.
Research (surveys): Survey definitions, questions, response data (may include personal data if respondents provide it), distribution links, completion timestamps, device type, country derived from the IP address, and precise coordinates only when the respondent's browser shares them.
Analytics: Aggregated metrics derived from survey responses. Account health scores, CLV predictions, and campaign models use organization-provided data. No additional personal data is collected.
AI tools: Written answers are analysed for themes, sentiment and summaries; the knowledge base indexes documents you upload; Synthetic Data generates fictional personas. This processing uses Groq (see Subprocessors). Comment translation runs on our own servers.
Platform roles (owner, admin, member, viewer) control what data each user can access. Row-level security in PostgreSQL ensures that users can only query data belonging to their organization. Role assignments are stored in our database and synchronized with Clerk metadata for session validation.
We use self-hosted analytics (on our own servers, not shared with third parties). Analytics tracking is only activated after you explicitly accept analytics cookies via the cookie consent banner. If you decline, no analytics data is collected.