Key Insight
People think in intentions — "make the sky dramatic" or "remove that person" — not sliders and masks. The interface needed to match how people describe what they want.
Design Direction
Built around natural language intent instead of manual tool selection. Instruction Based Editing shipped on S26 — text-guided manipulation. Sketch to Image turns rough drawings into realistic output.
Created a cross-cultural design framework for AI face features, now protocol at Samsung Seoul.
Design Tradeoffs
Full auto-generation was intentionally constrained. Users wanted AI assistance, not replacement. The design preserves human agency — the AI suggests, the human decides.
Outcome
2024 — Photo Assist on Flip6 and Fold7. 2025 — Instruction Editing on S26. 2026 — Roadmap extends to Fold8.
Generative UXPrompt FrameworkCross-Cultural AIBengaluru → Seoul
🔒 Some details under NDA. Happy to discuss in conversation.