Decoding Culinary Cultures
MenuDecode: The AI-Powered Dining Assistant
MenuDecode: The AI-Powered Dining Assistant
MenuDecode: The AI-Powered Dining Assistant
How I executed a 3-day rapid concept sprint using generative AI (V0 & Gemini) to build a psychology-led prototype, bridging the context gap for cross-cultural diners.
AI Concept Sprint
AI Concept Sprint
UX Psychology
UX Psychology
Product Strategy
Product Strategy
0-to-1 MVP
0-to-1 MVP
Executive Summary
Executive Summary
Executive Summary
The Challenge: International diners often face "Menu Anxiety" and decision paralysis when navigating text-only foreign menus, as literal translation apps fail to provide vital visual and cultural context.
The Strategy: Executed a 3-day, zero-budget rapid sprint using generative AI (V0 & Gemini) to design a psychology-led MVP. To ensure technical feasibility, I structured a tiered image-sourcing algorithm (fallback logic) rather than relying on a perfect, non-existent database.
The Outcome: Delivered an interactive React/Tailwind prototype that preserves the original menu layout to prevent AI hallucinations, provides empathetic visual context, and features a bespoke screen to facilitate the final offline ordering interaction with staff.
The Challenge: International diners often face "Menu Anxiety" and decision paralysis when navigating text-only foreign menus, as literal translation apps fail to provide vital visual and cultural context.
The Strategy: Executed a 3-day, zero-budget rapid sprint using generative AI (V0 & Gemini) to design a psychology-led MVP. To ensure technical feasibility, I structured a tiered image-sourcing algorithm (fallback logic) rather than relying on a perfect, non-existent database.
The Outcome: Delivered an interactive React/Tailwind prototype that preserves the original menu layout to prevent AI hallucinations, provides empathetic visual context, and features a bespoke screen to facilitate the final offline ordering interaction with staff.
The Challenge: International diners often face "Menu Anxiety" and decision paralysis when navigating text-only foreign menus, as literal translation apps fail to provide vital visual and cultural context.
The Strategy: Executed a 3-day, zero-budget rapid sprint using generative AI (V0 & Gemini) to design a psychology-led MVP. To ensure technical feasibility, I structured a tiered image-sourcing algorithm (fallback logic) rather than relying on a perfect, non-existent database.
The Outcome: Delivered an interactive React/Tailwind prototype that preserves the original menu layout to prevent AI hallucinations, provides empathetic visual context, and features a bespoke screen to facilitate the final offline ordering interaction with staff.
Context
Context
Context
Beyond Literal Translation
Beyond Literal Translation
Beyond Literal Translation
Experiencing authentic foreign cuisine often comes with a barrier: text-heavy, culturally specific menus. While standard translation tools (like Google Lens) convert words into the user's native language, they fail to provide the cultural and visual context required to make a confident dining choice. Diners are left frantically switching between translation apps, Google Images, and review sites—breaking the immersive dining experience.
Experiencing authentic foreign cuisine often comes with a barrier: text-heavy, culturally specific menus. While standard translation tools (like Google Lens) convert words into the user's native language, they fail to provide the cultural and visual context required to make a confident dining choice. Diners are left frantically switching between translation apps, Google Images, and review sites—breaking the immersive dining experience.
Experiencing authentic foreign cuisine often comes with a barrier: text-heavy, culturally specific menus. While standard translation tools (like Google Lens) convert words into the user's native language, they fail to provide the cultural and visual context required to make a confident dining choice. Diners are left frantically switching between translation apps, Google Images, and review sites—breaking the immersive dining experience.
The Psychological Insight
The Psychological Insight
The Psychological Insight
Empathy Over Algorithms
Empathy Over Algorithms
Empathy Over Algorithms
Leveraging my background in psychology, I framed this not merely as a language problem, but as a psychological one.
Cognitive Overload: Processing a dense menu whilst cross-referencing multiple apps overwhelms the user's working memory.
Neophobia & Decision Paralysis: The fear of the unknown often dictates choice. Without seeing what a dish looks like or understanding its cultural significance, diners retreat to "safe", unauthentic options, missing the true culinary experience.
Leveraging my background in psychology, I framed this not merely as a language problem, but as a psychological one.
Cognitive Overload: Processing a dense menu whilst cross-referencing multiple apps overwhelms the user's working memory.
Neophobia & Decision Paralysis: The fear of the unknown often dictates choice. Without seeing what a dish looks like or understanding its cultural significance, diners retreat to "safe", unauthentic options, missing the true culinary experience.
Leveraging my background in psychology, I framed this not merely as a language problem, but as a psychological one.
Cognitive Overload: Processing a dense menu whilst cross-referencing multiple apps overwhelms the user's working memory.
Neophobia & Decision Paralysis: The fear of the unknown often dictates choice. Without seeing what a dish looks like or understanding its cultural significance, diners retreat to "safe", unauthentic options, missing the true culinary experience.
Architecture
Architecture
Architecture
Structuring Trust: The Fallback Logic
Structuring Trust: The Fallback Logic
Structuring Trust: The Fallback Logic
As a product manager, I knew a concept reliant on a non-existent, perfect image database would fail a technical feasibility check. To ensure reliability without an established backend, I designed a tiered, auto-degrading visual verification system:
Priority A (Gold Standard): Merchant or user-verified uploads.
Priority B (High Relevance): Google Maps API matches (User photos tagged at the location with >75% text relevance).
Priority C (Generic Reference): Top Google Images matching the translated dish name (>75% text match).
Priority D (Fallback): Generative AI image, strictly labelled as an "AI Concept" based on ingredients, managing user expectations.
As a product manager, I knew a concept reliant on a non-existent, perfect image database would fail a technical feasibility check. To ensure reliability without an established backend, I designed a tiered, auto-degrading visual verification system:
Priority A (Gold Standard): Merchant or user-verified uploads.
Priority B (High Relevance): Google Maps API matches (User photos tagged at the location with >75% text relevance).
Priority C (Generic Reference): Top Google Images matching the translated dish name (>75% text match).
Priority D (Fallback): Generative AI image, strictly labelled as an "AI Concept" based on ingredients, managing user expectations.
As a product manager, I knew a concept reliant on a non-existent, perfect image database would fail a technical feasibility check. To ensure reliability without an established backend, I designed a tiered, auto-degrading visual verification system:
Priority A (Gold Standard): Merchant or user-verified uploads.
Priority B (High Relevance): Google Maps API matches (User photos tagged at the location with >75% text relevance).
Priority C (Generic Reference): Top Google Images matching the translated dish name (>75% text match).
Priority D (Fallback): Generative AI image, strictly labelled as an "AI Concept" based on ingredients, managing user expectations.
Execution
Execution
Execution
The "One-Person Team" AI Workflow
The "One-Person Team" AI Workflow
The "One-Person Team" AI Workflow
Operating on a zero-budget, 3-day sprint, I bypassed traditional wireframing. I utilized Gemini to outline the system logic and edge cases, and guided V0 via precise text prompts to instantly generate a high-fidelity React/Tailwind frontend. I applied Gestalt principles directly through prompt engineering to correct AI layout errors.
Operating on a zero-budget, 3-day sprint, I bypassed traditional wireframing. I utilized Gemini to outline the system logic and edge cases, and guided V0 via precise text prompts to instantly generate a high-fidelity React/Tailwind frontend. I applied Gestalt principles directly through prompt engineering to correct AI layout errors.
Operating on a zero-budget, 3-day sprint, I bypassed traditional wireframing. I utilized Gemini to outline the system logic and edge cases, and guided V0 via precise text prompts to instantly generate a high-fidelity React/Tailwind frontend. I applied Gestalt principles directly through prompt engineering to correct AI layout errors.
Designing for Reassurance
Designing for Reassurance
Designing for Reassurance
Every UI decision was purposefully crafted to build psychological safety:
Anti-Hallucination Structure: The scanner digitises the physical menu whilst strictly preserving original categories, ensuring users can always map the digital interface back to the physical paper.
Progressive Disclosure: To combat cognitive load, the initial list view is minimalist. Detailed macronutrients, ingredients, and visual verification are housed within specific dish modals.
Empathetic Microcopy: I designed an "AI Smart Picks" recommendation engine that uses reassuring cultural storytelling (e.g., "Don't fear the name!") to encourage culinary exploration.
Every UI decision was purposefully crafted to build psychological safety:
Anti-Hallucination Structure: The scanner digitises the physical menu whilst strictly preserving original categories, ensuring users can always map the digital interface back to the physical paper.
Progressive Disclosure: To combat cognitive load, the initial list view is minimalist. Detailed macronutrients, ingredients, and visual verification are housed within specific dish modals.
Empathetic Microcopy: I designed an "AI Smart Picks" recommendation engine that uses reassuring cultural storytelling (e.g., "Don't fear the name!") to encourage culinary exploration.
Every UI decision was purposefully crafted to build psychological safety:
Anti-Hallucination Structure: The scanner digitises the physical menu whilst strictly preserving original categories, ensuring users can always map the digital interface back to the physical paper.
Progressive Disclosure: To combat cognitive load, the initial list view is minimalist. Detailed macronutrients, ingredients, and visual verification are housed within specific dish modals.
Empathetic Microcopy: I designed an "AI Smart Picks" recommendation engine that uses reassuring cultural storytelling (e.g., "Don't fear the name!") to encourage culinary exploration.
The Final Touchpoint: Offline Integration
The Final Touchpoint: Offline Integration
The Final Touchpoint: Offline Integration
The digital journey inevitably ends with an offline human interaction. I designed a high-contrast, large-text, bilingual "Show to Waiter" summary screen, seamlessly facilitating the actual ordering process in noisy, high-pressure restaurant environments.
The digital journey inevitably ends with an offline human interaction. I designed a high-contrast, large-text, bilingual "Show to Waiter" summary screen, seamlessly facilitating the actual ordering process in noisy, high-pressure restaurant environments.
The digital journey inevitably ends with an offline human interaction. I designed a high-contrast, large-text, bilingual "Show to Waiter" summary screen, seamlessly facilitating the actual ordering process in noisy, high-pressure restaurant environments.
Future Strategy
Future Strategy
Future Strategy
Expanding the Ecosystem
Expanding the Ecosystem
Expanding the Ecosystem
While initially a B2C utility, the architecture is designed for broader commercial viability. The roadmap includes a social growth loop via "Share Bill/List" features (WhatsApp/QR) and a transition towards a B2B SaaS model, empowering local SMB restaurants to digitise their menus effortlessly.
While initially a B2C utility, the architecture is designed for broader commercial viability. The roadmap includes a social growth loop via "Share Bill/List" features (WhatsApp/QR) and a transition towards a B2B SaaS model, empowering local SMB restaurants to digitise their menus effortlessly.
While initially a B2C utility, the architecture is designed for broader commercial viability. The roadmap includes a social growth loop via "Share Bill/List" features (WhatsApp/QR) and a transition towards a B2B SaaS model, empowering local SMB restaurants to digitise their menus effortlessly.