Which workspace you are using
The connected local/staging workspace uses authenticated accounts and a private PostgreSQL database for your growth plans and evidence. The explicitly selected sample demo uses synthetic records and resets on reload. Sample demo interactions never call an AI provider.
This is a draft notice for a local/staging implementation. Production operator details, processor agreements, backup schedules, rights procedures and launch-market requirements must be completed before public launch. Use fictional work in testing.
Who is responsible
Kioni’s legal operator, business address, country of establishment, and privacy contact have not yet been designated. A working contact for privacy requests and any required EU or UK representative or data protection officer details must be published before public launch.
Until then, this is a draft for local/staging testing. Please use fictional details in the forms. Cookie and theme preferences can be cleared directly in your browser without contacting us.
Information and sources
The connected workspace stores the profile, permissions, growth areas, actions, work examples, corrections and career reports you save. Authentication uses Clerk for verified email, password and social sign-in, optional MFA, and session management. Kioni checks session status with Clerk before allowing private API access. Browser theme and cookie choices are separate from AI permission; see the cookie policy.
Connection information such as IP address, headers and request time reaches the hosting and network infrastructure. Production logging recipients and retention must be documented before launch.
Optional AI analysis
AI assistance is off by default. Enable it in workspace Settings or the analysis panel after reviewing the current policy. Enabling permission does not start a model request. Every analysis requires a separate Analyse selected context action; navigation, typing, saving work and task completion do not trigger analysis.
Your context preview identifies the saved evidence fields and optional growth cycle, milestones and tasks to be sent through Vercel AI Gateway to the configured approved model provider. Unselected evidence, reference URL contents, correction history and unsaved drafts are excluded. Model names and serving routes are controlled by the operator.
Requests require EU provider inference. Provider zero-data-retention routing is enabled by default; the operator can temporarily disable that requirement for testing. Gateway processing may occur outside the EU; EU inference is not an end-to-end EU residency guarantee. Provider and contractual arrangements must be qualified before real-work use.
AI output is provisional until the completed result and selected source references are checked. It may still be wrong. It offers descriptive observations and reflection questions, not competence scores, promotion decisions or automatic changes to your work.
Retention and deletion
Career reports reference only selected examples. Their text is composed from current sources when viewed or exported. Corrections require report review; deleting an example removes its excerpt from dependent reports. Export receipts expire after 24 hours and do not store copied report content. Your own narrative remains separate.
Unsaved analysis candidates expire after 24 hours. Saving creates a private analysis that remains until you delete it or a selected source is deleted. Source corrections mark analyses stale. Deleting a selected source removes dependent generated text, including saved analysis text.
Disabling AI cancels pending requests and removes unsaved candidates; saved analyses remain yours and stay readable after access expires. You can delete them individually. In-flight information already transmitted to a provider cannot be recalled, and stopping display does not guarantee immediate upstream cancellation.
Unsaved generation metadata is retained for up to 30 days, with conservative cost reservations for interrupted requests. Saved analyses retain source and model provenance; session and token-usage details are cleared after 30 days. Full prompts are not deliberately retained as logs. Production backup expiry and restore/deletion procedures remain launch prerequisites.
Your controls and requests
You can update profile and independent permissions in Settings, correct/delete individual evidence and delete saved analyses. Optional AI can be disabled without losing manual workflows. Account-wide portability and account deletion remain unavailable in this local/staging milestone and must be implemented before public launch.
Applicable privacy rights and request procedures depend on the final operator and launch markets. A monitored privacy contact, identity-verification procedure and required response/appeal processes must be published before public launch. Browser cookie choices do not authorize AI processing.
Security and automation
Ownership, permissions and saved changes are checked on the server. No website can guarantee complete security; avoid real confidential or third-party information in this testing environment. AI has no database tools and cannot fetch reference URLs or commit actions.
Kioni does not use this feature to make employment, competence or promotion decisions. Task completion and self-reported outcomes are not independent proof of competence.
Changes and contact
Material new AI purposes require a reviewed policy and appropriate acknowledgement before activation. Previously stored AI permission does not automatically enable this analysis policy.
Kioni’s legal operator, business address, country of establishment, and privacy contact have not yet been designated. A working contact for privacy requests and any required EU or UK representative or data protection officer details must be published before public launch.
Until then, this is a draft for local/staging testing. Please use fictional details in the forms. Cookie and theme preferences can be cleared directly in your browser without contacting us.