← RETURN TO DOSSIERFILE ELF.26 · CASE RECORD
EXHIBIT DPRODUCT · AI — CASE STUDY

SHIPSCORE

An AI-powered startup idea scorer — from product strategy and scoring logic to interface and deployment, turning uncertain ideas into ranked business decisions.

ORGANIZATION

INDEPENDENT PRODUCT

PERIOD

2025

ROLE

PRODUCT STRATEGY · SCORING LOGIC · UI · DEPLOYMENT — SOLO

STACK

REACT · CLAUDE API · NETLIFY

01 — THE BRIEF

Builders don't lack ideas — they lack an honest way to rank them. Gut feel is biased, friends are polite, and the cost of picking wrong is months of wasted build time.

ShipScore puts an idea through structured AI evaluation and returns a score a founder can compare and act on — the whole product designed, built, and shipped by one engineer.

02 — OPERATING CONSTRAINTS

  • An AI opinion is worthless if it's vague — scoring had to be structured and comparable across ideas.
  • Solo build: product strategy, scoring logic, UI, and deployment all one pair of hands.
  • Ship fast and cheap: a product experiment, not an infrastructure project.

03 — SYSTEM STRUCTURE

REACT INTERFACE

Idea in, verdict out — the scoring flow kept simple enough to use in a coffee break.

SCORING LOGIC

Structured evaluation criteria — the product's real IP — driving consistent, comparable scores.

CLAUDE API

The AI evaluator, constrained by the scoring rubric rather than free-styling opinions.

NETLIFY DEPLOYMENT

Zero-ops hosting so the experiment could ship and iterate immediately.

04 — KEY DECISIONS

DECISION 01

A rubric, not a chatbot

Free-form LLM feedback flatters and drifts. Fixed scoring criteria make two ideas comparable — which is the entire point of scoring.

DECISION 02

Own the whole loop

Strategy, logic, UI, deployment by one person means the product's thesis stays coherent from prompt to pixel.

DECISION 03

Deploy on zero-ops infrastructure

Netlify let the product exist in the world the day it worked — evidence beats a local demo.

05 — OUTCOME

Uncertain ideas became ranked business decisions.

  • · Second AI product in the file alongside NXRiseUp — different provider (Claude vs Gemini), same discipline: AI constrained by structure.

CROSS-EXAMINATION

The full walkthrough — code, schema, trade-offs — happens live. Fifteen minutes, no slides.

NEXT EXHIBIT

MARKET INTEL