The project involves building a text-based chatbot that can chat naturally with our clients and give reliable, friendly advice on general dog care. I want the bot to understand free-form questions (feeding schedules, grooming routines, exercise needs, basic wellness pointers, etc.), pull the right answer from a curated knowledge base, and reply instantly through a simple text interface on our website and, later, inside our mobile app. Core expectations • Conversational accuracy: the NLP must recognise common dog-related phrases, misspellings, and emotive language, then map them to the correct care tip. • Easy upkeep: I should be able to update or expand the knowledge base without touching code. • Smooth integration: embed the chat window in our existing React site and expose a REST endpoint so our app team can reuse the same engine. • User hand-off: when confidence drops below a threshold, the bot flags the chat and hands it to a human agent by email or Slack. • Analytics: basic dashboard or API that shows volume, top questions, and unresolved queries. Preferred stack is open—Python with Rasa, Dialogflow CX, or a lightweight Node solution are all fine as long as you explain the trade-offs. Model training data, deployment script (Docker or similar), and concise setup documentation are part of the final hand-over. A brief demo on your own server first, then we migrate to our AWS account. When the migrated bot matches the demo behaviour and the above criteria, the job is done.