AI Avatar Stylist
Designing a conversational AI tool that made avatar customization feel like play.
Context
The avatar editor is part of a product ecosystem with 1B+ avatars created — giving users a persistent digital identity across social apps and immersive VR. It's the central hub where users build and evolve their look.
Challenge
Avatar customization was powerful but overwhelming. Hundreds of options, nested menus, no guidance. Most users gave up before finding a look they loved. Meanwhile, AI was transforming creative tools — but no one had solved how conversational AI should work inside a 3D creative environment.
My role
Owned this project end-to-end — from the first brainstorm and prototype through public launch at scale. I defined the vision, scoped and planned every milestone, and orchestrated a multidisciplinary team across each phase.
- Defined interaction model, UX architecture, and design principles
- Scoped staged delivery: Prototype → Dogfood → Public Test → Launch
- Directed supporting designers, art directors, 3D artists, animators
- Partnered with brand team and CDs on AI naming and framing strategy
- Partnered with UXR on concept testing, usability, and longitudinal studies
- Presented strategy to leadership and 100+ person audiences
- Ran cross-functional office hours and creative reviews
Key design challenges
Setting expectations with AI
Users ranged from expecting magic to deep skepticism. I designed honest onboarding and 100+ curated starter prompts that teach capabilities through inspiration, not instruction.
Overcoming the blank canvas
A text field is intimidating. I created a contextual suggestion system with personality-driven prompts and follow-ups that maintain creative momentum without feeling prescriptive.
Options, not answers
A single AI output feels take-it-or-leave-it. I designed a multi-output "ingredients grid" where users see several options and can mix, refine, or iterate between rounds.
Bridging AI and manual control
Users want AI magic and fine-grained control. I designed seamless mode-switching between AI styling and the traditional editor — start with AI, hand-tune details, or vice versa.
Building brand trust
Early testing showed Meta AI branding led to over-expectations. I removed it, redesigned feedback loops, and set accurate expectations through design cues and copy.
Public reveal
Style with AI was unveiled on stage at the Meta Connect 2024 developer keynote — the public moment that set the stage for launch.
Impact
Drove 400K+ registrations to the Meta mobile gaming platform — 3rd largest growth contributor across all features.
Users spent 3 times more on paid content vs. traditional editing.
Ease-of-use satisfaction exceeded benchmarks.
Retention above platform benchmarks — users returned consistently.
Beyond launch
- Published 5 reusable AI UX patterns — adopted across other consumer AI teams
- Set the North Star vision for Prompt-to-Avatar (text/voice/photo → full avatar generation)
- Led 2026 AI Editing strategy planning with XFN team
- Onboarded and mentored designers, established design ops templates used team-wide
As featured in
0→1 Product Design
AI/LLM UX
Vision Setting
Cross-Functional Leadership
UX Research Partnership
Interaction Design
Design Ops
Prototyping
Mentorship