Family Elf · Product Management, Design & UX, Strategy
How Family Elf turned a founder-built prototype into a sprint-ready AI app
I joined Family Elf as product lead to turn a vibe-coded prototype into a scoped, documented product that a three-person engineering team could ship sprint by sprint.
- 84
- Features specified
- 53
- Roadmap items
- 21
- Features shipped

- Client
- Family Elf
- Services
- Product Management, Design & UX, Strategy
- Year
- 2026
- Tools
- Jira, GitHub, Clerk, Stripe, PostHog
- Engagement
- Fractional product lead, since June 2026
The company
Family Elf is a non-profit building a private AI helper for families. A household sets up Elves, AI helpers that each carry Skills like meal planning or medication reminders, and groups them into Projects with shared context. Every family member gets a role, from Owner to Child.
The challenge
One founder had vibe-coded most of the app and the marketing site. Auth, the household data model and what a Skill is technically were all undecided, and each one changed the database schema. Features had no rules for edge cases, and the live site sold a pricing plan the team had already dropped.
What I did
Set up how the team works
I wrote a project setup doc covering 19 tooling and infrastructure choices, each with a recommendation, alternatives and an owner. We landed on Jira, GitHub, Next.js, Clerk, Stripe and PostHog, three environments and a three-day standup rhythm.
Mapped the whole product
I built a product brief that walks every page, from sign-up to settings, as entry point and outcome. Each feature lists its dependencies, priority, open questions and the engineers' recommendation, so disagreements surfaced on paper.
Sequenced it into a roadmap
I grouped the brief into eight initiatives, from Auth and Onboarding to Chat and Super Admin, and slotted each feature into Now, Next, Later or Backlog with a MoSCoW priority.
Wrote specs engineers can build from
I created PRD, requirement and bug templates, then used them for the website redesign and the plans and billing work. Every PRD breaks down into an Epic, stories and front-end and back-end sub-tasks with Given/When/Then criteria.
Priced it like a non-profit
I designed Free and Starter plans capped by household messages per month, an invite-only founding cohort that stays free, and a cost-per-message guardrail tied to break-even, so the model covers costs without chasing margin.

Results
- 84
Features specified
Each with an entry point, outcome and open questions in the product brief
- 53
Roadmap items
Across 8 initiatives, prioritized from Must to Nice-to-Have
- 21
Features shipped
From auth and onboarding to Elves, Skills and Projects
Want results like these? Let's talk.
A 30-minute call about where your product is and what to do next.