RECOVERAI: A Context-Grounded WhatsApp Conversational Assistant for Personalized Post-Discharge Patient Follow-Up

Post-discharge care requires patients to follow medication instructions, attend appointments, understand recovery plans, and communicate concerns after leaving a hospital. This literature review examines digital follow-up, mobile medication-adherence interventions, healthcare conversational agents, and large language model (LLM) safety to identify design requirements for RecoverAI. The reviewed studies show that digital communication can support continuity and patient engagement, but effects on readmissions and adherence vary by intervention and population. Research on conversational agents also highlights the need to evaluate usability, information quality, safety, and clinical outcomes rather than treating message exchange as evidence of effectiveness. RecoverAI is an academic prototype that connects a React and Vite staff dashboard, Supabase PostgreSQL records, a Supabase Edge Function, an OpenWA WhatsApp gateway, and Gemini-assisted response generation. Demonstration records confirm retrieval of medication name, dosage, frequency, and instructions from a registered patient record, and delivery of an appointment reminder to a test phone. These observations establish selected integration paths only; they do not establish clinical benefit, complete medication-reminder scheduling, multilingual validation, or production readiness. The review identifies research needs in patient identity verification, grounded response accuracy, reliable message processing, accessible language support, privacy, and outcome-based evaluation.