The Smith School of Business at Queen’s University is part of a $20-million collaborative project to create a digital wellbeing platform designed to help people better understand their physical, mental, and emotional wellness. Under Canada’s Digital Technology Supercluster, Queen’s is working with lululemon, Wysdom.AI, Microsoft, and Mitacs on the project.

Wellbeing.ai will be an app that uses virtual agents, which will function like an advanced chatbot, to help users better understand their physical and mental health. Users will be able to speak and interact with the virtual agent, which will in turn provide personalized insights and support like recommending exercises, products, health resources, breathing techniques, and more. lululemon is looking to democratize wellbeing by making tools and resources available to more people, and Wellbeing.ai is part of this vision.

lululemon is the project’s lead private sector partner, with Wysdom.AI using their conversational AI-optimization technology to create the virtual agents, Microsoft developing a responsible AI ethical framework, and Mitacs supporting Queen’s in its academic research. The project has a tentative completion date of 2025.

Queen’s University is the academic partner on the project, and Stephen Thomas, Executive Director of Smith's Analytics and AI Ecosystem, is leading the university’s team of PhD students and post-docs. They are responsible for conducting research and experiments to ensure that the latest and best technology will be used in the creation of this wellness agent.

They will focus on four main areas:

• Intent recognition — the chatbot needs to understand what is being asked and said by users 
• Speech recognition — what users say needs to be decoded into text so that the chatbot can use the information 
• Virtual agent diagnostics — how the chatbot can automatically determine how well it is doing with users and fix itself or diagnose a trend when it frequently gets something wrong;
• Recommender systems — to provide recommendations to users that are not only accurate but personalized.

“I believe in the purpose of the project — the end goal of the project is to build this wellness resource for everyday citizens, to improve people’s day-to-day lives in a scalable way, to give them physical, mental, emotional help,” Dr. Thomas says. “If we do this right, it has the potential to be rolled out worldwide and help people who might otherwise not get the help.”

Queen’s will train the AI leaders of the future through access to industry funding and applied work on the Wellbeing.ai project.

“It’s going to offer incredible research opportunities for all of the PhD students working on the project. Top tier publications will come out of this, and they will get to work with industry partners on industry data sets on real world problems. It’s very practical and applied, yet very challenging at the same time,” Dr. Thomas adds.

Master of Management Analytics and Master of Management and AI students may also find themselves with snippets of these four main problems as class projects over the next few years. Dr. Thomas says that this will provide students with practical experience while also benefiting the project by having more eyes look at a problem.

“Queen’s Partnerships and Innovation (QPI) was crucial to getting this project up and running at the beginning. The Digital Technology Supercluster has a robust process to ensure that project partners and proposals meet all of their requirements,” he says.

QPI helped Queen’s build a relationship with lululemon and Wysdom.AI and assisted with the development of the project scope, milestones, budget, and the complex master project agreement.

“The goal of project is to build a virtual wellness agent to help people with their physical wellness, mental wellness, and emotional wellness,” Dr. Thomas says. “This has basically never been done before.”

For more information on the initiative, check out the website.