TOEIC Snack PM Case Study

AI-assisted product experiment | From personal pain point to MVP delivery and product tradeoffs

How a TOEIC reading-practice app moved from a personal pain point around fragmented study time into a timed-practice MVP, then used hands-on trials, iOS QA, a risk register, and AI-agent delivery mechanisms to validate product assumptions.

Period

3.5 months / about 25 implementation days

Role

Solo PM / AI Agent Orchestrator

Delivery

13 screens · 172 tests · iOS QA

Status

MVP demo ready; launch paused before business validation

PM Case StudyAI CollaborationReact NativeNestJSSupabase

MVP device screens: reading-practice core loop

The MVP supports practice entry, timed answering, result feedback, and history tracking.

Entry → Answer → Result → Retention

TOEIC Snack home screen showing level, streak, accuracy, daily target, and practice-duration choices — Practice entry
Level, streak, accuracy, and 5/10/15/20-minute choices turn fragmented study time into a ready-to-start task.

TOEIC Snack answer screen showing a Part 5 question, progress, timer, and four options — Timed answer
TOEIC part, question progress, countdown, and option states map directly to reading-speed training.

TOEIC Snack result screen showing accuracy, answer time, time bank, streak, and review — Result feedback
Accuracy, answer time, streak, and review are condensed into actionable feedback for one practice session.

TOEIC Snack history screen showing total sessions, weekly practice, average accuracy, and session records — History tracking
Session records, weekly filters, and average accuracy support return visits and continued practice motivation.

Case Brief

This case demonstrates product tradeoffs from a PM perspective, not only technical delivery.

I moved from a personal pain point to MVP scope, then used hands-on QA and handoff documents to control delivery quality.

PM Evidence

Narrowed a personal pain point into MVP scope
Controlled delivery quality with device QA, handoffs, and a risk register
Used AI agents to break down requirements and complete technical delivery
Judged when not to rush launch, preserving resources for higher-priority validation

screens

iOS-first MVP

172

tests

0 TS errors

QA bugs

All iOS QA findings fixed or triaged

PM capability this case is meant to prove

PM capability	Evidence in this case	Maturity
Problem definition	Moved from personal pain point, forum secondary research, and competitor scan into the angle of timed reading practice.	Proven practice
Scope narrowing	After completing the MVP, re-evaluated each feature by validation value; later projects now start with explicit non-goals.	Proven practice
Validation mindset	Five family/colleague trials, roughly 10 personal sessions, and iOS QA that found 11 issues.	Early validation
Risk judgment	Clearly listed next-stage outcome metrics, question-bank quality, and screen-level testing to add.	Proven practice
Technical collaboration	Decomposed requirements into work AI agents could execute, using handoff documents and QA to control delivery quality.	Proven practice

Core project insight

A product experiment that reached MVP, showing the full decision chain from vague pain point to hypothesis, MVP delivery, and next validation steps.