I need an external AI agent wired into Respond.io that greets every new enquiry, leads a natural chat to collect our booking basics—shoot type, preferred date, location—and then shares the right service package, all while drawing on an already-prepared knowledge base to answer follow-up questions. The moment a lead proposes an exact date or asks something outside the predefined scope, the conversation must be handed over to a human without losing context. You will be building a hybrid system: clear rule-based steps for qualification and escalation, reinforced by machine-learning NLP so the agent understands free-form questions and keeps the tone friendly and on-brand. All data gathered in the chat has to flow straight into our CRM, and any outgoing email (quotes, confirmations, package JPGs) should be triggered seamlessly from the same workflow. Social channels can stay out of scope for now. Deliverables I expect: • Fully functional AI agent connected to Respond.io, email, and (if possible) our CRM • Clean, well-commented source code plus setup scripts • Conversation templates and fallback rules encoded from the supplied knowledge base • Handover logic that creates an internal note and alerts the human team inside Respond.io • Short deployment guide so we can run, test, and iterate without you Once the above is in place and tested against a set of sample dialogues, I’ll consider the project complete and release the final milestone.