RAG-assistants for the Hospitality Industry

Leveraging Large Language Models to improve operations

Client: LosjiTech

Description

LosjiTech is a digital platform designed to manage bookings, issues, and other processes in the hospitality industry. To streamline access to the platform's extensive knowledge base containing documents related to hotel procedures and policies, we developed an AI-powered assistant. This assistant enables both hotel staff and guests to efficiently retrieve information.

For instance, the chatbot can answer questions like:

• What is the checkout procedure for Hotel X?
• What is the view like in room 123?
• Can I bring my dog?

We implemented this solution by embedding the relevant documents and storing them in a vector database, allowing the assistant to retrieve precise information based on user queries.

Impact

The assistant significantly improved the efficiency of information retrieval, enabling faster response times for both staff and customer inquiries. It has helped reduce the time spent on searching for procedures and policies, leading to enhanced productivity for hotel staff and a smoother user experience for guests.

Additionally, a separate chat interface was created for submitting issue forms, which further optimized operational processes by automatically filling certain fields, reducing manual input.

Tools & Methods

• Retrieval Augmented Generation (RAG)
  • RAG-evaluation
  • Corrective RAG (CRAG)
• Large Language Models
• Vector Databases
• Natural Language Processing (NLP)

Developed by Simula's Department of Applied AI.