AI Skin Analysis App UX
- UX lead
- Landscape analysis
- Approach & Proposal
- Feature UX
- User journey exploration
- Whitelabel definition
- MVP UX & protoyping
- UX lead
- Landscape analysis
- Approach & Proposal
- Feature UX
- User journey exploration
- Whitelabel definition
- MVP UX & protoyping
This idea started out as an exploration into machine learning and how we could use AI to create a product for our beauty & personal care clients which then evolved into an AI Skin analyser, quiz and recommendation engine taht can be used together or stand alone. I worked on a landscape analysis, feature ideation, proposals, whitelablel product MVP and prototype. This is now been rolled out to 6 global brands, going live in 18 markets.
I actually worked on a landscape analysis as part of another piece of work which focused on beauty and personal care diagnostics tools across 21 brands. Working in collaboration with the strategist we created a set of experience dimensions in combination with a Kano review to give us an experience score card to measure against when evaluating the market. These included:
Experience Dimensions- Brand personality
- Intuitiveness
- Personalisation
- Diagnostic quality
- Connected Experience
Kano Expert Review
- Fundamentals - Basics Essentials
- Expectations - Performance - Conscious comparators
- Exciters - Delighters - Shape and shift perception
Once we had compelted our landscape analysis, we had immersed ourselves in the context and could now provide guidelines on what to do and what not to do in order to create a great diagnostic experience. This then laddered up to us creating some overarching experience princples to be guided by when creating these experiences.
Now that we had the research and the guideling princples, we starting thinking about the MVP user journey. We decided to start by stripping the user journey right back to basics in order to focus on the flow in a more holistic way rather than getting too much into the detail of each screen. We looked at the quiz journey first to understand the most frictionless user flow through the screens.
We then explored features and functionality, what they could look like, how we could introduce micro animations to indicate behavours and how a user could interact with them ie. draging or swiping answers in order to make it easy and fun.
We wanted the tool to be able to be used by multiple brands and do multiple things so that it could work as modules that can be turned on or off, depending on the requirements.
I worked with the tea to define the modules and customisable layouts, creating wireframes and protoypes to demonstraight how different versions might work.
These included- A quiz flow with questions and answers
- the AI skin analysis with selfy upload
- Logged in profile experience
- Different layout options for landing screen
- The options to change icons, colours and branding
One of the first brands to test out our new product was the beauty brand AHC. They wanted the ability to have the user completed the AI skin analysis then log back in regularly to check and update their skin based on skin cycle information. So we created a 'beauty coach' experience for them, adapted all the branding, layouts and imagery and created a tailored log in experience that allowed the user to see the progress of their skins' improvements after using the AHC recommended products. Visit AHC beauty coach website
As of March 2021 we had priorised new features to add into our experience according to different brands needs. These included
- Consumer feedback loop
- Login, set and track goals
- 3P weather API
- Loyalty mechanism
- Generative content
- WhatsApp integration
- Influencer integration
- Recommendations by email
Results
The AI skin analyser/beauty coach is now a key product offering called AiBA, which has been rolled out for Unilever brands such as AHC, All, Dove, Glow & Lovely, LAKME and MELE skincare. The roadmap includes adding new features such as loyalty, weather API and WhatsApp integration. This is now been rolled out to 6 global brands, going live in 18 markets.
- 10 weeks to build MVP platform
- 6 global brands on board
- 18 markets
- 2 external partners integrated
- 74,207 diagnostics completed
More projects.
