AI-based planning became practical in the last few years. NutriWitty's plan engine uses Anthropic's Claude — not on autopilot, but inside well-defined guardrails.
Core question: can AI make medical decisions alone? No. In our system, every plan is a first draft by AI, then approved by our dietitian. Legally, KVKK article 6 on health data processing mandates this step.
AI is strong at: evaluating parameter diversity (age, height, weight, goal, allergy, diet preference, chronic condition, activity level, seasonal context) at once and producing coherent recommendations. A human dietitian can do the same — but it takes longer. AI cuts this from 45 minutes to 3.
AI is weak at: nuanced cases, rare medical conditions, motivational support. This is where the dietitian steps in. In NutriWitty chat the AI draws its line clearly: 'This should go to your dietitian — leaving a short note for them.'
Bottom line: the AI + human hybrid model beats both pure AI and pure human approaches. Speed (AI) + responsibility (human) = scalable, safe nutrition coaching.