UPDay (AI, B2C)
End-to-end Product Design to an AI-powered app.
Role
AI Product Designer
Industry
AI, B2C, Social Media
Duration
5 months
Situation
Upday was envisioned as an AI-powered social media app designed to connect health-conscious individuals based on shared habits, such as cycling, running, or trying vegan restaurants. Unlike existing platforms, which focused on either tracking activities or facilitating large group meetups, Upday aimed to bring deeper, one-on-one connections.
The challenge was to design an experience that nicely incorporated machine learning (ML) to recommend people, activities, and places that aligned with users' interests, all while maintaining a simple, engaging, and trustworthy user journey.
Task
My role as the product designer was to:
Conduct thorough research to understand users’ needs and frustrations.
Design an intuitive platform that allowed users to provide meaningful data for accurate ML recommendations.
Create a trustworthy and transparent interface that built confidence in the AI’s suggestions.
Collaborate closely with AI engineers to ensure the technology aligned with user expectations.
Action - Discovery Phase
I started with qualitative and quantitative research to define the problem and gather insights:
User Interviews:
I interviewed 20 individuals who led healthy lifestyles. They expressed frustration with the lack of platforms designed for building meaningful connections around shared habits.
Quantitative Surveys:
I distributed surveys to a wider audience, collecting data on popular healthy activities and preferences for social connections as well as analysing experiences and recordings from Hotjar.User Personas:
Based on the findings, I developed personas to represent our target users. For instance, “Emma” was a vegan runner looking for a buddy for weekend jogs and new restaurant experiences.Competitor Analysis:
I analysed similar platforms like Strava and Meetup. While these apps excelled in tracking and organising group events, they lacked features for fostering individual, interest-based connections.Defining the User Journey:
Using the research insights, I mapped out the user journey, highlighting key pain points and opportunities. The critical need was a personalised onboarding experience where users could share their interests to inform the AI’s recommendations.
Design and Development
Introducing the Interests Screen:
I designed a screen in the onboarding process where users could select their interests, such as “cycling,” “vegan restaurants,” or “yoga.” This information became the foundation for the ML system to recommend meaningful matches and activities.User Flows and Prototyping:
I created detailed user flows to ensure a smooth transition between discovering the app, setting preferences, and exploring recommendations. Low-fidelity prototypes were tested with real users to gather feedback.Iterating the Design:
Early tests revealed that the original interests screen was overwhelming, with too many options displayed at once. I simplified it by categorising activities into smaller, logical groups, making the process easier for users to complete.Building Trust in AI Recommendations:
Users were sceptical about AI’s accuracy. To address this, I added explanations to each suggestion, such as: “Recommended because you both enjoy cycling at similar times.” This transparency made users feel more confident in the app’s capabilities.Collaborating with AI Engineers:
I worked closely with the AI team to understand how the algorithm analysed user preferences and matched them with similar individuals. This collaboration ensured that the design supported the AI’s functionality while staying user-friendly.
Who doesn't love a little chaotic looking prototype?
Results & Impact
The project delivered impactful results such as:
39% increase in onboarding completion rates due to the simplified and engaging interests screen.
26% boost in user engagement, with users spending more time exploring recommendations and connecting with others.
85% of users formed a meaningful connection within their first month, as revealed through post-launch surveys.
Situation
Upday was envisioned as an AI-powered social media app designed to connect health-conscious individuals based on shared habits, such as cycling, running, or trying vegan restaurants. Unlike existing platforms, which focused on either tracking activities or facilitating large group meetups, Upday aimed to bring deeper, one-on-one connections.
The challenge was to design an experience that nicely incorporated machine learning (ML) to recommend people, activities, and places that aligned with users' interests, all while maintaining a simple, engaging, and trustworthy user journey.
Task
My role as the product designer was to:
Conduct thorough research to understand users’ needs and frustrations.
Design an intuitive platform that allowed users to provide meaningful data for accurate ML recommendations.
Create a trustworthy and transparent interface that built confidence in the AI’s suggestions.
Collaborate closely with AI engineers to ensure the technology aligned with user expectations.
Action - Discovery Phase
I started with qualitative and quantitative research to define the problem and gather insights:
User Interviews:
I interviewed 20 individuals who led healthy lifestyles. They expressed frustration with the lack of platforms designed for building meaningful connections around shared habits.
Quantitative Surveys:
I distributed surveys to a wider audience, collecting data on popular healthy activities and preferences for social connections as well as analysing experiences and recordings from Hotjar.User Personas:
Based on the findings, I developed personas to represent our target users. For instance, “Emma” was a vegan runner looking for a buddy for weekend jogs and new restaurant experiences.Competitor Analysis:
I analysed similar platforms like Strava and Meetup. While these apps excelled in tracking and organising group events, they lacked features for fostering individual, interest-based connections.Defining the User Journey:
Using the research insights, I mapped out the user journey, highlighting key pain points and opportunities. The critical need was a personalised onboarding experience where users could share their interests to inform the AI’s recommendations.
Design and Development
Introducing the Interests Screen:
I designed a screen in the onboarding process where users could select their interests, such as “cycling,” “vegan restaurants,” or “yoga.” This information became the foundation for the ML system to recommend meaningful matches and activities.User Flows and Prototyping:
I created detailed user flows to ensure a smooth transition between discovering the app, setting preferences, and exploring recommendations. Low-fidelity prototypes were tested with real users to gather feedback.Iterating the Design:
Early tests revealed that the original interests screen was overwhelming, with too many options displayed at once. I simplified it by categorising activities into smaller, logical groups, making the process easier for users to complete.Building Trust in AI Recommendations:
Users were sceptical about AI’s accuracy. To address this, I added explanations to each suggestion, such as: “Recommended because you both enjoy cycling at similar times.” This transparency made users feel more confident in the app’s capabilities.Collaborating with AI Engineers:
I worked closely with the AI team to understand how the algorithm analysed user preferences and matched them with similar individuals. This collaboration ensured that the design supported the AI’s functionality while staying user-friendly.
Who doesn't love a little chaotic looking prototype?
Results & Impact
The project delivered impactful results such as:
39% increase in onboarding completion rates due to the simplified and engaging interests screen.
26% boost in user engagement, with users spending more time exploring recommendations and connecting with others.
85% of users formed a meaningful connection within their first month, as revealed through post-launch surveys.
