AI-CRM-System: An Intelligent Customer Relationship Management System for Healthcare Professionals

AI-CRM-System is an open-source CRM system designed specifically for the healthcare industry. It integrates Groq LLM to provide real-time WebSocket chat agent functionality, supporting healthcare professional (HCP) management, interaction recording, automated workflows, and offering a complete Dockerized deployment solution. Key features include HCP profile and interaction management, a modern React frontend, and intelligent agent-driven business processes. It adopts a three-tier architecture: PostgreSQL data layer, FastAPI service layer, and React frontend layer.

The healthcare industry is a relationship-driven field, but traditional CRM systems fail to meet its special needs:

1. Strict compliance requirements: HCP interactions need detailed records for audit purposes;
2. Complex decision chains: Hospital procurement involves multiple stakeholders;
3. High-frequency interaction scenarios: Academic conferences, department meetings, one-on-one visits, etc.;
4. Personalized communication: Different specialists focus on significantly different information. AI-CRM-System provides an AI-integrated open-source solution to address these pain points.

Backend Architecture

Uses FastAPI framework, with dependencies including PostgreSQL (database), Groq API (LLM service), and WebSocket (real-time communication). The core data models are HCP (including id, name, specialty, and other extended fields) and Interaction (including id, hcp_id, interaction content, and timestamp).

AI Chat Agent

Achieves bidirectional real-time communication via the WebSocket /ws endpoint, supporting automatic reconnection and heartbeat; the agent can trigger tools (query HCP information, record interactions, generate reminders, etc.), parse natural language intentions, and execute corresponding operations.

Frontend Architecture

Based on React 18, supporting dark mode, typewriter effect, and memory function; tech stack includes React Hooks/Context (state management), Fetch/Axios (HTTP client), and native WebSocket API.

HCP Management

Full lifecycle management: Profiles (basic information, contact details, professional tags, relationship history), intelligent search (fuzzy matching, specialty/hospital filtering, interaction time sorting).

Interaction Recording and Tracking

Detailed recording of interaction types (visits, calls, etc.), topics, feedback, and attachments; provides a timeline view (display by time, filter search, highlight key events).

AI Chat Agent

Natural language interaction (query HCP dynamics, record interactions, generate reports, set reminders), context awareness (remember conversation context, associate HCPs, provide suggestions based on history), tool invocation (automatically call APIs, validate data, handle exceptions).

Automated Workflows

Follow-up reminders (generate tasks based on interaction time, priority sorting, multi-channel notifications), report generation (periodic activities, regional performance, compliance audits), data synchronization (external system integration, data cleaning, backup and archiving).

Docker Full-Stack Deployment (Recommended)

1. Copy .env.example to .env, set DB_PASSWORD and GROQ_API_KEY;
2. Execute docker compose up --build to start the service;
3. Access: Backend API (http://localhost:8000/docs), frontend interface (http://localhost:3000), WebSocket (ws://localhost:3000/ws). Service components: PostgreSQL (port 5432), FastAPI (8000), React (3000).

Local Development Mode

Backend: cd backend → pip install -r requirements.txt → uvicorn app.main:app --reload; Frontend: cd frontend → npm install → npm run dev (supports hot reloading).

Data Protection

PostgreSQL data encryption, sensitive information (API Key, passwords) managed via environment variables, Docker internal network isolation.

Audit Tracking

All data changes are logged, interaction records are tamper-proof, and data export is supported for audit purposes.

Access Control

Role-based permission management, API authentication and authorization, session management.

Compliance Recommendations

Implementation based on regional regulations: GDPR (EU), HIPAA (US), China Cybersecurity Law (data localization storage).

Current Limitations

Limited mobile support, multi-language support to be improved, advanced analytics (BI/prediction) to be developed, lack of standard integration with EHR systems.

Future Improvements

Mobile app (React Native/Flutter), offline mode, advanced analytics dashboard, FHIR standard integration, voice input, multi-tenant SaaS version.

Summary

AI-CRM-System integrates traditional CRM with AI capabilities to improve the efficiency of healthcare sales teams; Docker deployment lowers operation and maintenance thresholds, and its modular architecture facilitates customization and expansion. It is a solution worth evaluating for the healthcare industry and can also serve as a reference project for full-stack development.