Privacy Policy
Last updated: April 2026
What We Collect
- Assessment data: Your role, tasks, goals, and concerns (provided by you during intake)
- Career documents (optional): CVs and job descriptions you choose to upload or paste for the Full Diagnostic assessment
- Telegram user ID: For session management (not your phone number)
- Email: Only if you opt in via SEO pages
- Analytics: Anonymized usage events (no PII)
What We Don't Collect
- Phone numbers
- Location data
- Contact lists
- Browsing history outside our site
How We Use Your Data
- Generate your personalized AI resilience assessment
- Extract structured signals (tasks, skills, role context) from your uploaded documents using AI models. Every extracted item is grounded in evidence from your own text — we do not invent skills or tasks.
- Improve our AI models (anonymized, aggregated only)
- Send you follow-up emails (only if you opted in)
Your Rights
Under GDPR and similar data protection regulations, you have the following rights:
- Export: Use the
/mydatacommand in the bot to export all your data as JSON - Delete: Use the
/delete_mecommand to permanently remove all your data - Delete uploaded documents: Use the in-app deletion option, or
DELETE /api/privacy/delete-document/:id(theidis returned to you after each upload) - Email:Reply “unsubscribe” to stop email communications
Data Storage
- Assessment data: encrypted at rest in PostgreSQL
- Document text retention: Raw text from CV/JD uploads is retained for 24 hours for your review and re-extraction, then automatically deleted. Only structured signals (anonymized task/skill IDs and confidence scores) persist.
- Content hash: A hash of your uploaded document is kept for deduplication and support — never the text itself.
- Analytics: anonymized, hashed user IDs
- Industry reports: fully anonymized, aggregated
- Retention: non-document data is automatically deleted after 12 months of inactivity
Third Parties
- AI Models: Google Gemini, Anthropic Claude, Alibaba Qwen, and OpenRouter process your text for extraction and analysis. None retain your data after processing.
- Document Parsing: Marker (self-hosted, open-source) converts PDF/DOCX to text. Runs on our servers; no external data transfer.
- Telegram: message delivery only
- We do NOT sell your data to anyone
Cookies
We use minimal cookies strictly for functionality (e.g., consent preference). We do not use tracking cookies or third-party advertising cookies.
Contact
Questions about your data or this policy? Email: privacy@willai.app
See also: Terms of Service