← Back to Projects

Moodish

AI-powered meal recommendations based on how you feel

Moodish app interface showing mood-based food recommendations
01

Problem

Food delivery apps suffer from a key drop-off: users spend 8-10 minutes browsing before ordering, and 30% abandon without completing a purchase.

8-10 minutes browsing
30% abandon rate
02

Key Insight

Interviews with 11 users revealed a core mismatch between what apps offer and what users actually need.

Apps ask "What cuisine do you want?" while users think "I don't know what I'm in the mood for."

This disconnect revealed a deeper truth:

Food choices are driven by emotion, not logic. Users weren't asking for more options — they wanted guidance.
03

Solution

Moodish is an AI-powered app that recommends meals based on how you feel, not just past orders or cuisine categories.

Core Experience

01

Answer 4-5 quick multiple-choice questions about mood

02

AI agent delivers your meal recommendation with photo and personalized reasoning

03

Order directly through your preferred delivery app or discover nearby restaurants

See It In Action

04

Target Audience

  • Demographics: Adults ages 22-45 (millennials and older Gen Z)
  • Income: Middle to upper-middle income, regular food delivery users (2-4x per week)
  • Psychographics: Values convenience and personalization, decision-fatigued
  • Behavior: Frequently opens delivery apps but struggles to decide, spends 8-10+ minutes browsing
  • Pain Points:
    • Overwhelmed by too many food options and categories
    • Wastes time scrolling endlessly through menus
    • Apps focus on cuisine types when users don't know what they want
    • Seeks emotional connection to food but can't identify cravings
    • Frustrated by choice paralysis leading to order abandonment
05

Implementation

Collaborated with engineering team to implement AI personalization with cost constraints, using structured prompts to keep API costs under $0.15 per recommendation while maintaining quality.

$0.15 Cost per recommendation
3 min Reduced browsing time
06

Results

This approach reduced browsing time from 8-10 minutes to under 3 minutes, replacing decision fatigue with confidence.