Enterprise RAG Chatbot (internally configured as the Enterprise Knowledge Assistant) is a self-hostable Retrieval-Augmented Generation platform built for organizations that need employees to query internal documents — policies, engineering docs, HR handbooks, sales collateral — without leaking data across departments or paying for a dedicated on-prem GPU. It solves three problems that generic "chat with your PDF" demos usually ignore: ✅ Department-level data isolation — an HR employee's query never surfaces Engineering-only documents, because access filters are applied at the vector database query stage, before the LLM ever sees the retrieved chunks. ✅ No vendor lock-in on the LLM — the generation layer is provider-agnostic (OpenAI / Gemini / Groq), so you can run entirely on free-tier APIs (Gemini, Groq) or swap to OpenAI for production quality without touching the retrieval pipeline. ✅ Runs on modest hardware — embeddings are generated locally with a lightweight sentence-transformer model, so the only network calls are to the LLM API; no GPU, no 16GB+ RAM requirement.