Design and Deployment of a RASA CALM Driven WhatsApp Chatbot for Automated Customer Support

Conversational agents have progressed from simple rule-based responders to advanced dialog systems that understand context, user intent, and execute business-critical workflows. This paper describes the design, development and deployment of a RASA CALM (Conversational AI Language Model) architecture based WhatsApp customer-support chatbot deployed through Meta WhatsApp Business Cloud API. The chatbot automates product discovery, checks real-time stock availability, order related queries, multi-language support and human escalation. Integration with IMS, CRM system and ticketing engine were set. With defined requirements such as a sub-2 second response time, 99.9% uptime and scalable to have over 10,000 conversations concurrently all as stated in the technical specification document from the data we analysed, we see an 84.2% automated resolution rate, a human-escalation rate of only 9.7%, and CSAT score of 4.63/5 over 3,000+ users. We aim to draw a practical blueprint for realistic deployment of enterprise-wise AI customer service chatbots using open-source frameworks, production APIs.