Day One Retirement
AI-powered personalized retirement planning
Next.jsTypeScriptClaude APITailwind CSSVercel
- Timeline
- 4 weeks
- Role
- Solo PM/Engineer
- Status
- In Production
A comprehensive retirement planning tool that goes beyond finances to help retirees design fulfilling lifestyles. Through a 7-question personality quiz, the app generates personalized weekly schedules that match individual preferences and goals.
Impact
- •170+ unique visitors in the first 30 days
- •Conversion rate: ~9% for assessments and schedules built
- •~20% of converted users return within the next day
- •More research needed on retention features before proceeding past concept stage
Highlights
- •Conducted 5+ user interviews to validate problem space
- •Researched competitive personality tests & design a new scoring algorithm across 4 dimensions
- •Built AI prompt engineering system for schedule generation
- •Shipped 0→1 product to production in 2 weeks
Key Learnings
On AI Product Design:
- •Prompt engineering is product design: small wording changes can mean huge output differences
- •Cost per token matters when using AI coding tools. Optimization of plan to execute was my biggest unlock (LLM to Cursor)
- •AI outputs need validation: Built in manual review before showing schedules
On 0→1 Development:
- •Basic MVP iteration works: V1 had no weather integration but was easily added after getting feedback from users that results weren't local enough
- •5 users > 500 opinions: Deep interviews beat surveys for early insight
- •Time boxing works: 2-week constraint forced ruthless prioritization
What I'd Do Differently:
- •Start with simpler algorithm: The 4-dimension scoring was overkill initially
- •Mobile-first from day 1: 60% of users visited on mobile
- •Add analytics earlier: Wished I'd tracked funnel from the start
Screenshots
