AI WhatsApp Chatbot for Text, Voice, Images & PDFs with RAG

This workflow template consists of two interconnected automations for tech support teams: Document Ingestion & Indexing, and AI-Powered WhatsApp Query Response. The ingestion workflow manually imports product documentation from Google Docs, splits it into searchable chunks, generates OpenAI embeddings, and stores them in MongoDB Atlas for semantic search. The chatbot workflow listens for WhatsApp messages, processes text, transcribes voice notes, analyzes images, and parses PDFs, then retrieves relevant chunks via vector similarity search.

Leveraging GPT-4o-mini with RAG, it generates concise, context-aware answers while maintaining conversation history for multi-turn interactions. This eliminates manual doc searches, ensuring consistent, fast responses. Ideal for internal support, knowledge management, or customer service in tech, travel logistics, or any doc-heavy domain.

Benefits include massive time savings for agents, reduced errors from inconsistent answers, scalable knowledge base updates, and multimodal support for real-world queries like voice notes or screenshots. Use cases: Travel agencies answering driving regulations from PDFs, logistics firms querying route docs via WhatsApp, or product teams handling user manuals effortlessly.