wellnessai1

Wellbeing.ai

Understanding human interactions to deliver an immersive and personalized experience of wellness coaching.

Project Overview

Updated August 7, 2025

The Problem

The challenge with wellness coaching is in collecting the data needed to build out the full human model.

While physical health inputs like weight, activity, heart rate, hours of sleep are well defined and easily measured, that’s not the case with emotional and social health. This places a limit on being able to access conventional wellness coaching – generally one-on-one. Using machine learning and combined with multimodal data, this project will develop AI models and optimization products to understand human interactions and the environment to assess the state of wellbeing, overcoming a major hurdle in tailoring wellness management.

How We Are Solving It

lululemon athletica has partnered with Wsydom.AI, Queen’s University and Microsoft to lead a project powering a virtual agent that understands human interactions and environment to provide meaningful and targeted fitness and wellbeing engagement and encouragement.

This project will develop a model to understand wellbeing and human interactions to deliver an immersive and personalized experience of coaching. In the end, individuals will be able to work with a digital wellness agent to combine assessment and coaching to achieve their person wellness goals which will enable lululemon’s goal of providing access to wellbeing tools for 10 million people.

The Result

The Wellbeing.ai project successfully explored how virtual agents (VAs) could interpret human interactions and environmental context to assess human wellbeing and deliver personalized engagement. To overcome the complexity of optimizing a VA’s “digital brain”, the project consortium developed Wysdom’s Operations Center (OC), a proprietary analytics dashboard that provides deep insights into virtual agent performance. The OC uses AI to identify conversation topics, interaction volume, containment, conversation time and user feedback — enabling continuous, data-driven optimization for customer engagement.

lululemon implemented Wysdom’s OC in areas of its business operations to generate deep insights into their guest feedback and how lululemon can better serve them. From the time of implementation to 2025, Wysdom OC has increased lululemon’s Guest Satisfaction Score (GSAT) by +144% over three years.

The success of the project highlights the potential of AI to enhance customer experiences and provide personalized engagement to support wellbeing-focused applications in the future.

Project Lead

  • lulu e1633460713956

Project Partners

  • wysdom2 e1633471130553
  • queensweb e1633497909817
  • microsoft@2x e1632716841590
  • mitacs@2x e1632685086233

Get involved now!

We’re not waiting for the future. We’re creating it.

Become a Member

Are you ready to solve society's toughest challenges?