AI-powered chatbot makes critical information accessible across languages for bus drivers

Vy needed to make important information from a comprehensive handbook accessible to bus drivers in a busy work environment.

With more than 40 different languages represented among drivers, significant language barries arose. In addition, some drivers had reading and writing difficulties, making traditional documents less suitable.

The information was stored in a long and difficult-to-navigate document, making it time-consuming to find answers to specific questions. This affected efficiency, quality, and safety in daily operations.

Knowit, together with Vy, developed an AI-powered chatbot that provides bus drivers with fast and easy access to relevant information from the handbook.

The solution uses Retrieval Augmented Generation (RAG), which combines language models with document-based search. This ensures that responses are based on verified sources and reduces the risk of misinformation.

The chatbot supports both text and speech, allowing users to ask questions in their own language. Responses are presented in a clear and understandable way, tailored to each user.

Technologically, the solution is built on Databricks and AWS, using services such as AWS Bedrock, Cognito, and Transcribe. It is developed as an LLMOps setup with versioning, logging, and access control, making it robut and scalable.

The work was carried out in cross-functional teams using agile methods, where continuos testing and iteration ensured high quality and relevance.

The solution has med it significantly easier for bus drivers to find the right information when they need it. By removing language barriers and offering both voice and text interfaces, it has contributed to greater inclusion and better accessibility in daily work.

At the same time, it has reduced the time spent searching for information and made it easier to make the right decisions in operational situations. The solution was also used to train new bus drivers in the summer of 2025.

The project demonstrates how generative AI can help solve concrete challenges related to communication, efficiency, and workplace inclusion.