CASE STUDY
Provence en Couleur AI-Guide
Designing an AI-powered product guide for personalized retail experience.
100%
Data Automation
$5
Monthly Cost
Company: Provence en Couleur
Industry: Wellness & Health
Location: Vancouver, Canada
Business Model: Retail & E-commerce
Timeframe: September 2025 - Present
Role: AI Web Developer, UX\UI Designer, UX Researcher
Tools: Claude, Figma, HTML\CSS, OpenAI, and Google Workplace.
Background
Provence En Couleur is a Granville Island boutique specializing in French lavender and aromatherapy products. Seeing the potential to enhance their customer experience, I proposed creating an AI-powered product guide to help shoppers discover items aligned with their wellness goals and scent preferences. With the owner’s support, we began exploring ways to combine automation, personalization, and design clarity to elevate both in-store and online interactions.
Identified Problem
Shoppers are curious and eager to explore, but the abundance of choices and limited guidance—especially online—make it difficult for them to find products that truly fit their needs.
-
Visitors are intrigued but unsure where to start.
-
The wide product range can feel overwhelming.
-
Staff expertise isn’t easily scalable.
-
Lack of guidance online
UX Research & Insights
During in-store observations, I noticed that visitors often lingered; testing scents, chatting with staff, and asking questions like “What’s good for sleep?” or “What helps with stress?” These behaviors revealed a sense of curiosity and exploration rather than a clear purchase intent.
To better understand their motivations, I conducted five in-store interviews, which confirmed that most shoppers were driven by curiosity and a desire to learn rather than to make an immediate purchase
Solution
Building on this insight, I realized that AI could help bridge that moment of uncertainty by offering personalized guidance.
This led to the design of an AI-powered product recommendation system that invites visitors to Find Your Perfect Gift.
By answering just three simple questions, shoppers receive tailored product suggestions with clear prices and descriptions.
Although initially designed for in-store use, the same system can easily be adapted for the website or other digital touchpoints.
Building UX with AI →
Step-by-step process showing how AI tools like ChatGPT, Claude, and OpenAI APIs were integrated into a real UX workflow—from data structuring to design, testing, and machine learning refinement.
01
Catalogue Structuring
-
Used ChatGPT-5 to extract and organize the product catalogue (names, prices, scents, benefits).
-
Hosted structured data in Google Sheets as a live database.
-
Created a GPT as mockup. See it here.
02
AI Development
-
Used Claude for code generation and guidance in Wix Studio.
-
Integrated OpenAI API to analyze user inputs and match products dynamically at a cost of 5$ per month.
-
Built backend logic in JSW to manage API calls and filtering.
03
UI Design & Front-End
-
Applied the brand’s design system; optimized for kiosk use (large touch targets, clear progress).
-
Adjusted fonts and colours for accessibility under retail lighting.
04
QA and Machine Learning Maintenance
-
Logged mismatches and refined AI prompts.
-
Refined rules (budget limits, scent pairings) to improve accuracy.
-
Implemented idle-reset and offline fallback.
User Flow
The flow was intentionally designed to feel simple, calm, and intuitive; mirroring the sensory ease of the in-store experience. Each step reduces decision fatigue while progressively personalizing recommendations.
-
The user comes to the kiosk.
-
Chooses their budget.
-
Picks their preferred scents.
-
Selects their wellness goal.
-
Gets personalized advice and recommendations.
-
Informs the employee and collects their products.
AI Product Suggestion Results
This sample bundle was generated by the app to demonstrate how the system personalizes recommendations. Each item is selected through rule-based logic that interprets the user’s inputs—such as budget, scent preference, and wellness goal—to build a cohesive, tailored product set.
Potential & Impact
The project continues to grow, and it’s exciting to see how AI is unlocking new UX possibilities for the brand.
-
Empowers local business operations through AI-driven product logic.
-
Boosts in-store and online conversions.
-
Helps users and staff better understand the product catalog.
-
Builds a scalable foundation for future features like multilingual support, occasion-based recommendations, and inventory insights.