What it is
A 5-page workbook on building a recommendation engine that learns from negative signals. Covers the full live architecture powering Suggested for you on wellnessand.ai: dismissal capture, affinity decay with 14-day half-life, anonymous-to-authenticated handoff via the claim_session_dismissals RPC, server-side scoring weights, instant card replacement, three-layer caching with versioned invalidation, and the six analytics events that prove it works (cache hit/miss rate, dismissal rate, replacement latency, re-suggestion guardrail, anonymous-to-auth handoff rate). Includes code snippets, SQL health-check query, and a rollout checklist.
Who it's for
For practitioners scaling 1:1 work without scaling burnout.
Best for
- You run a coaching, nutrition, functional-medicine or longevity practice.
- You want to use AI as a delivery layer — not replace your clinical judgement.
- You're done re-explaining the same protocol in every consult.
Not the right fit if
- You want a fully managed software product — this is a build-it-yourself system.
- You're not ready to put your method in writing.
Audience: Practitioners · Domain: ai
How to use it
🔒 The step-by-step guide unlocks after purchase.