What traditional scouting is good at
Context. A scout in the stand knows the pitch was unplayable, that the player had been ill that week, and what the dressing room thinks of him.
Character and environment — the things no dataset carries.
Relationships that get a deal done once the decision is made.
Where it runs out
Coverage. A department can only watch the leagues it staffs. Every market it does not cover is a market where its competitors sign cheaper.
Recall. A player watched in March and not shortlisted is functionally forgotten by August.
Consistency. Two scouts describing the same player rarely produce comparable documents.
What an AI-assisted index adds
Continuous scoring: every published profile is matched against every open brief, so a player who becomes relevant when your brief changes surfaces on his own.
Comparable scoring across profiles, so a shortlist can be read side by side.
Speed at the top of the funnel — turning thousands of profiles into a watchlist of twenty in minutes instead of weeks.
Where AI-assisted scouting must not be trusted alone
It cannot judge temperament, injury history nuance, or whether a player will settle in a new country.
It reflects the data it is given. A profile with thin footage will be scored on thin evidence.
The right shape is: automation narrows, humans decide. Anyone selling the reverse is selling something.