S26 Style Studio

50+ AI styles proposed. 10 shipped globally

Project Goal

Design an AI portrait styling system that transforms photos into expressive artistic styles while preserving the subject’s identity, enabling users to explore creative visual transformations they can still recognize as themselves.

Role

AI style exploration, AI-Prompt testing, global cultural research, identity-aware UX design

Duration

3 months (2025) l Core team : Bengaluru → Seoul

Impact

10 styles shipped globally on Galaxy devices
Scalable prompt testing framework

Problem

Despite the rapid growth of generative AI image tools, creating expressive image styling still requires complex prompts and experimentation.

Users often struggle to achieve visually appealing results while maintaining personal identity.

Official Replay | Galaxy Unpacked February 2026 | Samsung

00:00/00:00

Approach

Researched emerging visual trends and future directions in generative AI image styling

Conducted 1-on-1 interviews and feedback surveys to understand how different demographics respond to various portrait styles.

Prioritized styles using a structured priortisation matrix to balance visual diversity.

Performed iterative prompt testing across Samsung and Google AI engines to refine style outputs and improve consistency.

Key Insight

People care about recognizability more than artistic fidelity.

When a style looked beautiful but users couldn't recognise themselves, trust in the AI dropped significantly.

Design Trade-Off's

Hyper-realistic styles were deprioritized.

When outputs became too photorealistic, users felt uncomfortable with AI-generated faces. Stylized outputs built more trust.

Extreme color grading was reduced.

While vibrant palettes looked visually striking, they sometimes overpowered facial details and reduced recognizability.

Cultural stereotypes were carefully reviewed.

Certain stylistic representations risked reinforcing clichés across regions, so globally inclusive aesthetics were prioritized.

Note

The styles shown here are part of the final curated experience.

In total, 50+ visual styles were explored, going through multiple rounds of experimentation, prompt testing, and internal critiques before the final set was selected.

01 / 12

Due to NDA compliance, only selected parts of the case study are presented here.

🔐 View full case study

S26 Style Studio

Project Goal

Role

Duration

Impact

Problem

Despite the rapid growth of generative AI image tools, creating expressive image styling still requires complex prompts and experimentation.

Users often struggle to achieve visually appealing results while maintaining personal identity.

Approach

Researched emerging visual trends and future directions in generative AI image styling

Conducted 1-on-1 interviews and feedback surveys to understand how different demographics respond to various portrait styles.

Prioritized styles using a structured priortisation matrix to balance visual diversity.

Performed iterative prompt testing across Samsung and Google AI engines to refine style outputs and improve consistency.

Key Insight

People care about recognizability more than artistic fidelity.

When a style looked beautiful but users couldn't recognise themselves, trust in the AI dropped significantly.

Design Trade-Off's

Note

The styles shown here are part of the final curated experience.

In total, 50+ visual styles were explored, going through multiple rounds of experimentation, prompt testing, and internal critiques before the final set was selected.

Final launched Style & Categories

Illustrative Styles: Watercolor, Sketch, Webtoon, Storybook, Oil Painting.

Character & Playful Styles: 3D Cartoon, Clay Figurine, Felt Toy, Peekaboo, Bubble Time.

Photography & Camera Effects: Studio Portrait, Golden Hour, Fisheye Lens, Sneaky Peek, K-Pop Star.

S26 Style Studio

Project Goal

Role

Duration

Impact

Problem

Despite the rapid growth of generative AI image tools, creating expressive image styling still requires complex prompts and experimentation. Users often struggle to achieve visually appealing results while maintaining personal identity.

Approach

Researched emerging visual trends and future directions in generative AI image styling

Conducted 1-on-1 interviews and feedback surveys to understand how different demographics respond to various portrait styles.

Prioritized styles using a structured priortisation matrix to balance visual diversity.

Performed iterative prompt testing across Samsung and Google AI engines to refine style outputs and improve consistency.

Key Insight

People care about recognizability more than artistic fidelity. When a style looked beautiful but users couldn't recognise themselves, trust in the AI dropped significantly.

Design Trade-Off's

Note

The styles shown here are part of the final curated experience. In total, 50+ visual styles were explored, going through multiple rounds of experimentation, prompt testing, and internal critiques before the final set was selected.

Final launched Style & Categories

Illustrative Styles: Watercolor, Sketch, Webtoon, Storybook, Oil Painting.

Character & Playful Styles: 3D Cartoon, Clay Figurine, Felt Toy, Peekaboo, Bubble Time.

Photography & Camera Effects: Studio Portrait, Golden Hour, Fisheye Lens, Sneaky Peek, K-Pop Star.

Despite the rapid growth of generative AI image tools, creating expressive image styling still requires complex prompts and experimentation.

Users often struggle to achieve visually appealing results while maintaining personal identity.

People care about recognizability more than artistic fidelity.

When a style looked beautiful but users couldn't recognise themselves, trust in the AI dropped significantly.

The styles shown here are part of the final curated experience.

In total, 50+ visual styles were explored, going through multiple rounds of experimentation, prompt testing, and internal critiques before the final set was selected.