← back
WhatsApp LLM based AI Support Agent Automation for Chatting with cusotmers

WhatsApp LLM based AI Support Agent Automation for Chatting with cusotmers

Pending
💰 INR 1500–12500 👤 Unknown 🕒 20d ago status: new
Artificial Intelligence OpenAI ChatGPT Generative AI Large Language Model GPT-4 AI Chatbot Development Conversational AI AI Model Integration Gemini
(Please bid only after reading all the requirements in detail. I will consider your bid amount as the final amount for the project. I am not interested in further negotiations. I will pay only through freelancer, dont expect direct payments. I will check your past work on a similar project. I will select u only if you have done any project on the Automation of Chats with LLM. I will need a public link to test the same.) Please build a fully-automated WhatsApp chat agent that connects to a Large Language Model and follows an agent framework capable of enforcing conversation logic and a friendly, on-brand tone. The agent’s purpose is customer support, with a focus on on-demand information sharing. In practice it must recognise incoming requests, decide whether the user wants product details, service updates or answers to our FAQs, and reply instantly with accurate, context-aware language. • Integrating the chosen LLM (GPT-4, Claude, or similar) through a middleware layer. • Designing an agent architecture that can: – store and recall conversation context, – follow predefined tone/voice rules, – chain internal tools or functions when extra data is needed, – Hand it off to a human using specific keywords or confidence thresholds. • Exposing simple hooks (REST or WebSocket) so future micro-services can push fresh product or service data without a full redeploy. • Logging every exchange for later analytics while keeping PII masked. Acceptance criteria • A Docker-based project that runs locally with sample env variables for API keys. • Demonstrate three scripted user journeys (product detail, service update, FAQ) recorded via WhatsApp sandbox. • Source code, brief setup instructions, and a concise “tone & logic” prompt configuration file so new team members can tweak the brand voice without touching code. Only developers who have shipped similar LLM chat automations before should take this on; efficiency and reliability matter more than experimentation. This doc has more details of the project https://docs.google.com/document/d/1If_ZW91RfYAFFbAZCVYVfstSOGFY82hL6o7WHuhHvlo/edit?usp=sharing project details
↗ View on Freelancer