I want to deploy an AI-driven automaton dedicated to customer support, acting as a round-the-clock assistant for my business. The bot should greet users naturally, understand their intent, pull answers from a knowledge base I will provide, and, when needed, hand the conversation off to a human agent without losing context. Your job is to design, train, and integrate this assistant so it can: • Handle the full span of routine support requests that come through our website chat and email • Recognise when a question falls outside the knowledge base and trigger escalation • Log every interaction to our existing CRM so we keep a complete customer history • Allow me to update FAQs or add new replies without touching code I’m open to a framework you’re comfortable with—Dialogflow, Rasa, LangChain, direct OpenAI API, or a comparable NLP stack—as long as it delivers reliable intent detection and easy maintenance. Provide the trained model (or prompt workflows), connection scripts, and short documentation that lets my team deploy and update the bot independently. When you finish, I’ll test by sending a batch of real customer queries. I’ll sign off once the bot answers accurately, escalates correctly, and we can see the transcript in our CRM.