The biggest challenge was creating a corpus which would train a model to answer a vast number/diversity of queries. Another problem was connecting the web application to the rasa server which we could not solve due to time constraints.