About This Architecture

SalesBoost is a multi-layer sales automation platform combining Next.js frontend, Node.js backend, Supabase data layer, and AI-driven lead intelligence. Traffic flows from sales users through WAF and CDN to a React dashboard, then to an API Gateway that routes requests to lead management, workflow orchestration, and messaging services. The Intelligence Layer uses LLM endpoints to classify leads as hot/warm/cold, generate workflows, analyze intent, and optimize campaigns via a vector database. External integrations with WhatsApp, Instagram, and Facebook Messenger feed inbound leads through webhook receivers back into the API Gateway, creating a closed-loop automation cycle. This architecture demonstrates event-driven workflow automation with real-time AI enrichment, essential for sales teams scaling lead qualification without manual overhead. Fork and customize this diagram on Diagrams.so to adapt the integration layer, add new AI models, or swap Supabase for your preferred database. The separation of presentation, application, data, and intelligence layers enables independent scaling of each tier.

People also ask

How do you architect a sales automation platform that uses AI to classify leads and automate workflows across WhatsApp, Instagram, and Facebook?

SalesBoost uses a four-layer architecture: a Next.js React frontend with workflow builder UI, a Node.js API Gateway routing to lead management and workflow engines, a Supabase data layer storing users/leads/workflows/messages, and an Intelligence Layer with LLM endpoints that classify leads as hot/warm/cold and generate optimized workflows. Inbound leads from WhatsApp/Instagram/Facebook flow throu