AI-Proof Job Scanner

AI-Proof Job Scanner is a web app that analyzes a job’s resilience against AI automation. In 30 seconds, you get a numerical score, an impact timeline, and concrete recommendations — not generalities.

How it works

You fill in three fields: your job title, your sector, and your main tasks. GPT-4o-mini analyzes this data by comparing it against automation patterns documented in recent economic literature.

The result includes:

• A resilience score from 0 to 100 (0 = automatable immediately, 100 = protected long-term)
• An impact timeline: when the first partial automations arrive, when the role is fundamentally transformed
• Personalized recommendations: skills to develop, pivots to consider, parts of the job to reinforce

The analysis accounts for dimensions that classical models ignore: task variability, required human interaction level, judgment complexity, and the regulatory context of the sector.

Architecture

Frontend — Next.js 14

• App Router with Server Components
• Single-step form, results displayed via streaming
• Deployed on Vercel (Hobby plan, free)

Backend — Next.js API Route

• Call to GPT-4o-mini with a structured prompt
• Response streamed via ReadableStream for progressive display
• No database — stateless by design

Prompt engineering

• System prompt with economic context (MIT, McKinsey, OECD studies on automation)
• Output forced into structured JSON: score, timeline, recommendations
• Low temperature (0.3) for consistent, reproducible responses

Why GPT-4o-mini and not GPT-4o?

The cost per request is 15x lower. For a free app in beta, this is the obvious decision. Analysis quality for this type of structured task is sufficient — GPT-4o-mini excels at classification and scoring tasks with a good prompt.

If the app moves to a paid version with deeper analysis, GPT-4o will be used for premium tiers.

Beta status and roadmap

The app is functional and publicly accessible. Next steps:

• Pre-analyzed job database (cache of the 500 most-searched jobs)
• Comparison with sector average
• PDF report export
• API version for integration into other HR tools

This project is open source — contributions welcome.

Built by Thomas Silliard — in beta since March 2026.