AI Use Case Scorer
Is this actually a job for AI?
Describe an AI use case. The scorer asks a clarifying question or two if it needs to, searches the web for tools that already exist, produces a four-axis score, and sketches a quick wireframe of what the app might look like.
Four axes, one verdict.
Real, painful problem people care about?
Not already available off the shelf?
Realistic ROI — time, money, decisions?
Buildable v1 with available tools and data?
Weighted 30 / 10 / 30 / 30. Composite is 0–100. Higher is always better. The model returns the four axis scores; the server enforces the math so the verdict can't be jailbroken into a higher number.
See the tool in action.
Strong candidate — build a scoped pilot first.
A Monday-morning brief for the head of sales: pulls the weekly pipeline movement from Salesforce, highlights deals that slipped, names the three reps who need attention, and ships as a slide in their inbox at 7am.
See the breakdown →Already free in your email client — skip or pivot.
An AI tool that summarizes long emails into one-paragraph briefs so busy professionals can triage their inbox faster.
See the breakdown →Tempting demo, treacherous in production.
Let the finance team ask natural-language questions of the data warehouse and get accurate answers without writing SQL.
See the breakdown →Right problem, wrong project for a small team.
A real-time legal advice agent for small business owners — answers questions about contracts, employment law, and compliance in plain English with cited sources.
See the breakdown →Scoring uses claude-sonnet-4.6 via OpenRouter, EXA for the web search, and OpenAI for the wireframe sketch and content moderation. Scored ideas persist via Vercel Blob at unguessable URLs.