Automatable Judgment Study

What it is

A research project asking whether the arrival of cheap, typed, calibrated machine judgment opens a durable business in commercial intermediation such as freight brokerage. The analytic framework was fixed before any data, then populated with 18 blind synthetic practitioner interviews and seven web-verified research briefs, and the resulting candidate products (C01 to C15) were attacked by six adversarial panels. The finding was that feasibility is abundant and incentive alignment is scarce, with two pre-registered hypotheses revised and one refuted.

The problem

Picking a business idea around a new AI capability usually rewards a good story rather than a measured map of the opportunity space.

How it works

1. METHOD.md and map/FRAMEWORK.md pre-register nine transaction-cost sources, a seven-part capability gate and scoring dimensions; amendments are appended and dated, never edited in.
2. A persona protocol and instrument drive blind synthetic practitioner interviews, with transcripts kept.
3. Research briefs R1 to R7 cover freight brokerage, adjacent intermediaries, the model's technical limits, competitors, and liability and pricing.
4. A static dashboard over JSON findings browses candidates, with panel write-ups under tests/.

Stack

Disciplines

For your business

Structured, pre-registered AI opportunity research that tests which processes in a client's industry are worth automating.