I am ready to add an AI-driven chat agent to my existing web application. The goal is simple: visitors should be able to type questions and receive natural, helpful answers in real time. Here is what I need from you: • Design and build the text-based chat agent, powered by a proven NLP engine (OpenAI, LangChain, Rasa, or something equally capable). • Integrate it seamlessly into my current web stack so users can start a conversation from any page without refresh. • Allow easy fine-tuning with my own FAQ, product docs, or future data sources so the agent’s responses stay on brand. • Provide a lightweight admin panel—nothing fancy, just a place where I can review logs, update knowledge, and toggle response settings. • Deliver clean, well-commented code plus setup instructions so I can deploy to staging and production on my own servers. Acceptance criteria 1. The chat widget loads in under two seconds on desktop and mobile browsers. 2. It answers at least 80 % of test queries correctly during user acceptance testing. 3. All dependencies and environment variables are documented in a README. 4. Handover includes a short video walkthrough or live demo call. If you have practical experience with real-time web sockets, vector databases for retrieval-augmented generation, or fine-tuning large language models, let’s talk—I’m keen to see how you can bring this agent to life.