The challenge
Reservation teams receive a constant stream of guest emails about availability, bookings and policies. Answering manually is slow, and the information is spread across documents.
Our solution
We built a Python engine with LangChain that ingests documents into a vector store and answers incoming emails with retrieval-augmented generation using OpenAI and IBM watsonx models, connected to the conversation platform as an email channel.
What we delivered
- Document ingestion pipeline
- Vector search with LangChain
- RAG answering with OpenAI and watsonx
- Email channel integration
The impact
- Faster, consistent replies to guests
- Answers grounded in the business's own documents