The project brings together several practical AI components into a single, production-ready solution that will live on both our website and a companion mobile app. At the front end, I need conversational chatbots that can effortlessly handle customer support, automate routine business workflows, and guide visitors through everything from simple FAQs to complex troubleshooting steps. Behind the scenes, those same bots will draw answers from a Retrieval-Augmented Generation (RAG) pipeline capable of parsing PDFs, Word documents, and existing knowledge bases. OCR is essential as well, with reliable Telugu and English text extraction feeding directly into the knowledge workflow. Every document that enters the system should be analysed, summarised, and indexed in a vector database (FAISS, Pinecone, or a comparable store) so the chatbots and any auxiliary search tools can surface precise, context-aware responses. Key deliverables • Deployed chatbot(s) for website and mobile endpoints, wired for customer support, business automation, and general site interaction • End-to-end RAG service that ingests PDFs, Word files, and KB articles, incorporating OCR (Telugu & English) where needed • AI-driven summarisation, sentiment analysis, and text-classification utilities exposed as reusable Python functions or micro-services • Automation scripts/agents written in Python to remove repetitive back-office tasks • Web interface or dashboards built with FastAPI, Flask, or Streamlit for administration and testing • Clean documentation outlining setup, API keys, environment variables, and how to extend the system with new data sources or LLM providers Acceptance snapshot The solution is considered complete when a user can upload a mixed batch of documents, watch the content appear in search within minutes, and see the chatbot answer relevant questions—equally well on the website or mobile app—while sentiment analysis and summarisation APIs return consistent, verifiable results. Preferred toolset references (flexible): LangChain or LlamaIndex for orchestration, OpenAI or Anthropic models for generation, and standard Python libraries for task automation. Continuous deployment to a cloud host (AWS, GCP, or Azure) is a plus, but I am open to suggestions that fit the overall architecture.