Development and Pilot Evaluation of an Evidence-Based Personalized Nutrition and Nutraceutical Guidance Chatbot in Pakistan: A Case Study

Background: Pakistan's nutraceutical market is expanding rapidly amid weak regulatory oversight and limited consumer access to validated guidance. Self-medication, mega-dose supplement use, and misinformation are prevalent. Currently, no consumer-facing tool provides personalized, evidence-based nutrition support. Objective: To develop and pilot-test an AI chatbot delivering evidence-based nutritional and nutraceutical guidance within strict non-diagnostic boundaries. Methods: Data Curation: Information sourced from NIH ODS, Medscape, UpToDate, and peer-reviewed literature. Product Evaluation: Assessment of 1,000+ supplements from 30+ companies against GMP, ISO, and DRAP label-claimed certifications. Personalization Logic: Adaptive, context-aware recommendation framework. Safety Protocols: Cumulative toxicity prevention, consent-based nutraceutical guidance, and explicit non-diagnostic disclaimers. Key Features: Non-Diagnostic Scope: Clear boundary setting with referral to healthcare professionals when appropriate. Label-Verified Database: Products screened within NIH ODS safe intake ranges. On-Request Supplement Module: Nutraceutical recommendations provided only upon user request. Results: Internal pilot evaluation demonstrated high information clarity (85% positive), system helpfulness (70% positive), and cautious but constructive user reception to automated product guidance. No safety boundary violations were observed. Conclusion: AI-driven, evidence-based nutrition guidance is feasible in developing-country settings with limited regulatory oversight. This model offers scalable consumer education to reduce supplement misuse and address health literacy gaps.