A chat bot that can speak to users like a human and help them find services.
About the project
The Chatbot developed by the Waterloo Region Innovation Lab is a system based on artificial intelligence (AI) technology, which is able to achieve basic conversations with users and provide the smart searching system for multiple services around the region. Specifically, through the Natural Language Processing technology, the Chatbot is able to analyze the topics, emotions behind what the user is saying and provide them with corresponding services information as well as necessary support. Moreover, 27 commonly used language types in the Waterloo Region are being supported to make the Chatbot more user-friendly.
The problem statement
Currently, most of the searching functions on websites can only support accurate keywords searching in limited language types (English/French). Users have to input the specific service name or its corresponding category to get the information they are looking for. However, hardly can users get any search result if they try to search using some “emotional phrases”. The lack of understanding common emotional search phrases and limited language support may cause difficulties for users, especially in such a multicultural country like Canada. With such motivations, our lab started to develop the Chatbot which is able to “understand” human’s feelings and provide a more user-friendly searching system.
We tried to apply our Chatbot as an add-on system to some existing accessing services, like Impact Map, Family Compass and One List, so that users are able to get services information just through the conversations with the Chatbot. Related staff for these accessing services all mentioned that the multi-language smart searching system provided by the Chatbot can really help with users from different backgrounds!
The AI Chatbot System could be really practical and applicable to various use cases. All users will be able to get support easily no matter what cultural background they are in.
Communication with different groups and service staff is significant since this makes us truly understand what is lacking and what is needed.
If we could time-travel and start this project again...
- Add on more supported language types not only based on what is commonly-used but also about what is really needed
- Develop the system in a way that is more user-friendly according to different user groups
Thank you to Hisy, Grayson, Saba, and the entire SWRIL team! A special thanks goes to Zac, Hargun, and Elizabeth for helping to support! We appreciate all of your hard work!
Currently, we are continuing to enhance the Chatbot system – its functionalities and supported language types- based on the comments from various kinds of organizations and services. If you are also having some suggestions to help us improve, contact us!