Working Well
- A clear, step-by-step shade quiz.
- Preferences beyond color and undertone.
- Results with accessible descriptions.
SkinIQ
Generative AI in Retail 路 Student Capstone
SkinIQ: AI Shade Finder
SkinIQ looks beyond color alone鈥攃ombining a shopper鈥檚 current routine, desired finish, skin type, season, and additional context to create a more personal makeup-matching experience.
The Problem
A color match is only part of a good makeup match.
Many online beauty tools focus mainly on skin color or tone. A shopper can still receive a foundation with the wrong finish or a formula that does not suit dry, oily, normal, or combination skin.
The Solution
Use the shopper鈥檚 full routine as context.
SkinIQ guides the customer through a short quiz about products already used, desired finish, skin type, season, and any additional details, then returns recommendations with plain-language descriptions.
Before and With SkinIQ
From color-only matching to context-aware recommendations
Traditional shade finder
- Centers the decision on shade or skin tone.
- Can inherit an error from a product the shopper already wears.
- May overlook finish, formulation, skin type, and season.
- Leaves the customer to compare many products manually.
SkinIQ experience
- Starts with the shopper鈥檚 real beauty routine.
- Captures matte, natural, or glowy/dewy finish preferences.
- Includes skin type, season, and open-ended context.
- Explains why each result may fit the shopper鈥檚 needs.
How SkinIQ Works
A guided path from routine to recommendation
1
Start the quiz
Begin the personalized shade-finding flow.
2
Add the current routine
Share the foundation or concealer currently being used.
3
Choose finish and skin type
Select matte, natural, or glowy/dewy and describe the skin type.
4
Add season and details
Choose year-round, summer, or winter and include extra context.
5
Review explained results
Compare recommended products and understand the match rationale.
Journey to the Idea
Assignment 4 revealed a gap in existing shade finders.
Ava and Maddie found that consumers increasingly expect personalized beauty shopping, yet inaccurate online purchases and color-only tools still create friction. Their research reframed the problem: shoppers care about how makeup looks, feels, and performs鈥攏ot shade accuracy alone.
Beauty is not one-shade-fits-all, and personalization is becoming the new standard.
Prototype Status
What is working, what is being refined, and what comes next
Still Being Refined
- A shopper鈥檚 current shade may already be incorrect.
- A color match may still have the wrong finish or formulation.
- Subtle undertones and screen-rendered swatches can be misread.
What鈥檚 Next
- Stronger confidence and validation signals.
- Richer product attributes across brands.
- A feedback loop that learns from real-world wear.
The Central Idea
A useful recommendation should match how the product will be worn.
鈥淪kinIQ turns shade finding into a more complete beauty decision鈥攐ne that considers color, finish, skin type, season, and routine together.鈥
Built in Five Sessions
From an overwhelming new technology to a working product
Across SA国际传媒鈥檚 five-session Generative AI in Retail workshop, the team moved from learning unfamiliar AI concepts on day one to building and presenting a working SkinIQ product in the final session. Talks from leaders at Google, OpenAI, Wayfair, and Observe.AI were paired with applied assignments that helped the idea become a browser-based experience.
Prototype Notice
Use recommendations as guidance, not certainty.
SkinIQ is a student prototype. AI-generated product names, descriptions, shade matches, price ranges, and analyses are for demonstration only and are not verified inventory, pricing, availability, or an official retailer integration. Screen appearance, lighting, and the shopper鈥檚 starting product may affect results.
Acknowledgment
Thank you.
SkinIQ was built by Ava Pitts and Maddie Lee through SA国际传媒 University鈥檚 Generative AI in Retail program, with guidance from Prof. Kirthi Kalyanam and the Retail Management Institute.