Skip to content
Back to dink.one

AI Policy

Last updated 2026-05-14. This page sits next to Privacy - it covers only the AI-specific bits that didn't fit cleanly in the privacy policy.

What our AI features do

Advisor. Weekly performance recap, social recap, and wellness check using your match results, ratings, and any opt-in fitness data. Read-only output - never sends a message on your behalf.

Scout. Pre-match opponent breakdown using public match history and ratings, generated when you tap "Get a scout report" on a confirmed booking.

Coach Q&A. Conversational answers to skill / strategy questions you ask explicitly.

Match-debrief notes. Optional summary of a match generated only when you opt in at the end of score entry.

What we never do

  • We never train third-party models on your data.
  • We never send your real name, phone, or email to a model provider.
  • We never auto-post AI-generated text to your messages, social accounts, or any other external surface without you explicitly tapping "Post".
  • We never use AI to make eligibility or moderation decisions for you.

What we send to the model

Inputs are anonymized server-side before leaving Supabase: participant display name > handle slug, phone/email stripped, address replaced with city + state. Output never travels back without passing through our anonymization layer.

Providers

Today: Anthropic (Claude). Provider may change; the sub-processor list in the Privacy Policy is the authoritative current state.

Your choices

AI features require an explicit one-time consent at first use. After that, you can revoke consent at any time in Settings > Privacy. Revoking deletes any stored AI-generated content tied to your account within 30 days.

Errors and hallucinations

AI output can be wrong. Treat advisor recaps, scout reports, and coach answers as starting points, not facts. Don't use them to make rating, booking, or payment decisions that aren't already supported by the underlying data we show alongside.