OVERVIEW
Designing a simple and intuitive AI-powered platform to support artisans in their digital communication.
93% of Craftsmen
Lack time for digital communication
Marketplace → AI Assistant
Strategic pivot driven by field research
End-to-end
From UX research to hi-fi prototype & branding
PROBLEM & RESEARCH QUESTION
CONTEXT
Artis is an AI-powered web application helping artisans promote their work on social media.
Initiated as part of the Gobelins Paris startup program, the project pivoted from a marketplace concept to an AI content assistant, enabling text, image, and hashtag generation with automated scheduling.
The goal was to allow artisans to showcase their craftsmanship without sacrificing time or energy on digital communication, through a simple, intuitive, and accessible solution.
PROBLEM STATEMENT
27 out of 32 craftsmen surveyed lacked time for social media. 22 struggled with product promotion. Budget and digital skills were the two other major blockers. Artisans often lack intuitive and accessible tools to manage their digital communication efficiently resulting in inconsistent content creation, wasted time, and difficulty maintaining a clear digital presence.
Research-Driven Design
Grounded in field interviews and a quantitative survey, this project was driven by a documented gap between craftsmen's real communication needs and the tools available to them. Research findings directly shaped every product decision, including the strategic pivot from marketplace to AI assistant.
RESEARCH GOALS
PERSONA & METRICS

Guillaume
Time-constrained
Craft-proud
Digitally ambitious
Frustrations
No time to post
Lacks budget
Poor visibility
Can't afford an agency
Based on
Interviews
Personas
Craftsmen surveyed
No time for content creation
AI generates content in minutes, no expertise needed
Doesn't know what to say online
Craft-specific prompts guide the content creation
Can't afford a communication agency
Affordable self-managed AI tool
KEY INSIGHTS & DECISIONS
HOW I APPROACHED THIS PROJECT
Based on field research and co-design workshops, we followed a structured Product Design process from discovery to high-fidelity prototype.
DURATION
4 Months
Limits
School project
No live deployment
No development
CASE STUDY - SOLUTION
From field research to go-to-market strategy, Artis is a full product design project covering brand identity, AI-assisted content generation, and a high-fidelity prototype built for craftsmen with no digital expertise.
Deliverable
Marketing Plan & Roadmap
Figma Prototype & Branding
Commercial Strategy
Expected Outcome
Less time on content
AI generates copy, visuals and hashtags in minutes
Craft-centered experience
Personalized guidance built around each artisan
All-in-one affordable tool
Create, schedule and publish across a single tool
Key Learnings
End-to-end is demanding
Covering different fields simultaneously requires strong project management
Research drives product decisions
The pivot from marketplace to AI assistant came entirely from field data
Constraints sharpen focus
We prioritized what actually mattered for craftsmen
If I started over
Start with a real user in the loop
Recruiting a craftsman as a project partner to help on every design and strategy decision
Involve craftsmen in the branding
Bringing real artisans in the project would have made it more authentically theirs
Test the AI flow earlier in the process
Prototyping the AI logic from the start would have revealed technical constraints sooner
What I keep
Research over intuition
Having 9 interviews and 32 survey responses as a constant reference kept every product decision grounded
Strategic pivot ownership
Defending a full product repositioning with field data taught me how to justify product decisions to stakeholders
Multi-disciplinary approach
Combining UX, branding and go-to-market in one project built a complete product vision
What comes next
Explore monetization model
Validate the commercial strategy with early adopters to confirm pricing and acquisition channels
Run usability tests on the prototype
Validate the AI content flow with real craftsmen before any development investment
Pilot with real artisans
Deploy a beta with 2-3 volunteer craftsmen to collect real usage data














