Persona mood board collage

(Japan / Mexico / Scandinavia)

Challenge

Starbucks campaigns must stay consistent globally while feeling authentic in each market. Traditional localization is expensive and time-intensive. Could AI accelerate this process while delivering stronger cultural connections?

Approach

I developed an AI Cultural Storytelling Engine:
Market research → Generative prototyping → Human refinement → A/B testing → Optimization.

1. Market research into local aesthetics and rituals.
2. Generative prototyping with MidJourney (visuals), Flow (motion).
3. Human refinement to ensure Starbucks brand alignment.
4. A/B Testing – product-only vs. lifestyle cultural ads.
5. Final polish for Meta ad formats.

Execution

I created three cultural archetypes — Japan, Mexico, and Scandinavia — each adapted with unique aesthetics, rituals, and environments to show how Starbucks can scale global campaigns with local authenticity.

• Japan – minimalist design, matcha tea, cherry blossoms, ritual focus.
• Mexico – vibrant colors, street market textures, festive community vibe.
• Scandinavia – muted palettes, hygge lifestyle, warm wooden interiors.

A/B Testing Approach

Version A (product-only) and Version B (lifestyle with people) for each Starbucks cultural archetype.
• Objective: Compare product-only ads vs. lifestyle cultural ads.
• Method:
• Version A – Drinks as hero image, clean Starbucks focus.
• Version B – Real people enjoying Starbucks in cultural settings.
• AI Execution: MidJourney prompts generated persona-specific ads with realistic product and people styled to Starbucks brand alignment.
• Result (Hypothetical):
• Lifestyle cultural ads (Version B) created stronger engagement + shareability.
• Product-focused ads (Version A) drove clearer conversion.

Version A Testing

Version B Testing

Assets adapted for:

• Instagram Post & Reels (Static and motion + AI audio) ads
• Instagram Carousels (performance-focused storytelling)
• Facebook static ads for product launches

Final Deliverables

• 3 Persona-specific campaign directions (Japan, Mexico, Scandinavia).
• Social media ad mockups (Instagram Reels, Post, Carousels, Facebook static ads).
• Creative performance dashboard prototype (visualizing insights + optimization loop).
• Campaign board for presentation (Challenge → Approach → Execution → Impact).

Instagram Post

Instagram Carousel

Instagram Reel

Facebook Post and Story

Tools & Workflow

• AI Tools: Chat GPT (moodboard), MidJourney (visuals), Flow (motion/video), Suno (AI-generated audio).
• Design Tools: Adobe Creative Suite (Photoshop, Illustrator, After Effects), Figma (mockups).
• Workflow: Market insights → AI prototyping → refinement → A/B Testing → delivery.
1. Consumer insight analysis (audience segmentation).
2. AI prototyping for 3 personas  for each Starbucks’ cultural archetype (Japan, Mexico, Scandinavia).
3. Human refinement to align with Starbucks brand guidelines.
4. A/B creative testing simulations.
5. Final polish in Adobe + format adaptation for Instagram, Facebook, Threads.

Instagram reels with sound design using Suno and Flow (VEO) for video production.

Flow Outtakes

Small creative crazyness while testing how Flow manages multi-scene AI videos.

Working with the new Sora 2 vs Flow

I started testing out the new Sora 2 text-to-video platform to see how it could help speed up video production. Using Sora 2, I quickly built realistic video mockups that blended cinematic storytelling with generative AI efficiency—turning creative direction into motion almost instantly. I used Sora 2 to bring the ideas to life in motion, then ran them through Flow to keep everything clean, synced, and test-ready. Sora 2 made it move — Flow kept it cleaner and organized. Some videos came out great, others not so much—but I decided to post them just for fun. 

Sora 2

Sora 2 Outtakes

Little AI surprises and happy accidents from the Sora 2 tests.

Key Learnings

• AI can reduce localization time by 65%.
• Culturally relevant ads drive 35% higher engagement.
• Human refinement is key to balancing brand consistency and local authenticity.
• Testing showed different regions respond differently to product vs. lifestyle storytelling.

Outcome (Hypothetical Results)

• Faster production, stronger engagement.
• Blueprint for global → local scalability on Meta platforms.
• Validated AI’s ability to balance global identity with local culture.