Request to develop architecture for a large AI Healthcare Chatbot

We are looking to develop an AI-powered healthcare chatbot to streamline patient interaction, preliminary triaging, and administrative workflows. Below are the core objectives, functional features, and non-negotiable technical requirements for the project. Please review these details and provide architecture of the project along with timeline, and cost estimate. 1. Core Purpose & Target Audience Primary Goal: Provide patients with a 24/7 conversational interface for basic symptom triaging, appointment booking, FAQs, and medication reminders, while reducing manual load on administrative staff. Target Users: Patients, clinical staff, and administrative team members. 2. Functional Requirements & Key Features AI Symptom Checker & Preliminary Triaging: Natural Language Processing (NLP) to understand patient-described symptoms. Guided triaging logic (categorizing urgency: Emergency, Urgent Care, Routine, or Self-Care). Strict safety disclaimers and emergency rerouting (e.g., immediate trigger to call local emergency services if red-flag symptoms like chest pain are detected). Appointment Scheduling & Management: Real-time slot lookup, booking, rescheduling, and cancellations. Automated SMS/WhatsApp/Email confirmations and reminders. Prescription & Medication Support: Medication dosage reminders and refill request processing. General drug interaction and intake guidance FAQs. Clinic & Administrative FAQs: Automated answers for operating hours, location directions, insurance coverage, and billing queries. Seamless Human Handoff: Smart escalation to live support staff or triage nurses when the query exceeds AI capability or involves sensitive medical issues. 3. Non-Negotiable Technical & Security Requirements Regulatory & Data Compliance: Strict compliance with HIPAA / GDPR / local health data privacy regulations. End-to-end encryption for data at rest (AES-256) and in transit (TLS 1.3). Integration Capabilities: Integration with existing Electronic Health Record (EHR) / Electronic Medical Record (EMR) systems via HL7 / FHIR standards. Payment gateway integration for consultation co-pays or booking fees. AI Guardrails & Accuracy Control: Retrieval-Augmented Generation (RAG) architecture trained strictly on verified medical databases (e.g., PubMed, ICD-10 datasets) to prevent AI hallucinations. Role-Based Access Control (RBAC) to ensure strict patient privacy. 4. Proposed Phased Approach Phase 1 (MVP): Administrative FAQs, basic appointment booking, and emergency-route triaging with basic EHR sync. Phase 2: Advanced AI symptom checking, automated prescription refills, multi-language support, and deep EMR integration. Next Steps & Discussion Points Please provide a response addressing the following: Relevant Portfolio: Examples of HIPAA-compliant healthcare applications or AI agents you have previously built. Rough Estimate: Initial timeline and budgetary estimate for Phase 1 (MVP). Tech Stack Recommendations: Your preferred architecture for security and scalability. Looking forward to your response. Best regards,