Case Study: Eddy Travels achieves 96% NLU accuracy and richer conversational experience with Rasa

Eddy Travels' AI-powered assistant increased the NLU accuracy to 96% and provided a much-improved chatting experience

Eddy Travels is a Toronto-based travel tech startup that built an AI-powered assistant to deliver personalized flight, hotel, and car rental suggestions across messaging apps. Initially the team used Dialogflow but ran into limits—dialogs were restricted to predefined flows, they had no source-code access, and the platform was hard to maintain or test, hindering collaboration and improvements.

Eddy Travels migrated to Rasa’s open-source platform, customizing the code and machine-learning models to gain full data control and handle more complex language. Using Rasa, the assistant’s NLU accuracy rose to 96%, the chat experience improved, and the bot now processes thousands of conversations per month with higher engagement and more searches per conversation.

Eddy Travels

Edmundas Balčikonis

Co-Founder and Chief Executive Officer