Savi: Eat Well. Spend Wisely. (Concept)

Concept project: a solo, end-to-end sprint exploring how generative AI can speed up product strategy and UX design, from research through prototype and pitch.

Healthy, affordable home cooking is out of reach for many households, because matching nutritious recipes to what's on sale that week is genuinely hard, and no existing app solves this end to end.

I evaluated three approaches: a recipe app, a deal aggregator, and an AI platform connecting weekly store deals to personalized meal plans. Competitive analysis confirmed the third was an unaddressed market gap.

The AI-first approach brought real risks: hallucinated data, inconsistent grocery feeds, and token costs at scale. So the process prioritized validating onboarding and the core value proposition before investing in back-end infrastructure.

Research: brand voice, design brief, and personas (working parents, budget students, retirees) with ChatGPT. Market analysis with Claude.

Naming: a phonaesthetic study with Claude led to the name and tagline "Savi: Eat well. Spend wisely."

Design: logo and moodboard explorations with Gemini, ChatGPT, and Figma. Brand style documentation with Claude. Five UI iterations in Figma Make and Figma.

Prototype: a semi-functional React prototype built with Figma Make, GitHub, and Vercel. It covers onboarding edge cases, a smart grocery list, over-budget meal plans, and "Fix my plan".

Test: a pilot usability test plan built with Claude and run on the onboarding flow.

A named brand (Savi), a React prototype, a usability-tested onboarding flow, and an investor pitch deck, built on a compressed timeline using generative AI across research, design, and testing.

Run the market-analysis survey with real users, validate regional grocery feed availability, and define the data model needed to connect AI-generated meal suggestions to live store inventory.

Full case study and prototype: https://sites.google.com/view/port4ward/work/eat-well-spend-wisely