AI decides
Low stakes, easy to reverse — data is enough
- Recommendation system
- Scheduling, routing
- Automatic translation
- Content editing
Iterators.org · Tools
AI fluency does not mean you can code. It means you recognise when AI creates real value — and when it does not. These three tools give you a framework for that.
Note: the matrices below are in Hungarian. The downloadable versions are Hungarian as well.
Interference topic · Where AI decides, and where the human does
What if the failure isn’t in the AI — but in delegating irreversible decisions to a place that needed human context?
01 — Decision Zone
Two axes: reversibility × context requirement. The three near-miss cases — pricing, customer prioritisation, recruitment — all belonged in the top-left, yet were delegated to AI.
High stakes, hard to reverse — only a human can supply the context
AI recommends, human decides — human supplies the context
AI decides, human supervises — data is enough, but oversight is needed
Low stakes, easy to reverse — data is enough
02 — AI Value Decider
Two axes: runs without humans × creates high value. Most organisations invest in "AI Noise" and "Waste" — while the "AI Gold" zone remains almost untouched.
High value, but doesn't work without human involvement
High value, works without humans — this is scalable AI
Low value, yet ties up human resources
Runs without humans — but who needs it?
03 — AI Fluency Self-Assessment
Two axes: do you recognise AI output × do you evaluate it critically. The goal is the top-right — but most people are in the bottom-left without knowing it.
You recognise it, but don't evaluate it — leaving value on the table
You know when to trust it — and when not to
AI decides for you — without you knowing
Recognises it, but doesn't evaluate — accepts without checking