Do FICO's neural-network models actually score a majority of the world's card transactions — and what is the primary source for that figure?
The claim claim-hnc-falcon-fraud-manager-became-fico-infrastructure carries a headline reach statistic that currently rests only on a Tier-3 venue (Machine Learning Times / Predictive Analytics World):
"Neural network models created by FICO score a majority of both the credit and debit card transactions in the world today."
Per the sourcing floor (00-meta/sources.md), any specific quantitative claim —
"a majority," "two-thirds," an "Nx" share — requires a Tier 1–2 primary. This
one does not clear the floor and is flagged [unverified-quant — needs primary].
Why it matters
It is the load-bearing payoff of the whole "connectionist commercial exit" thread: if true, a written-off 1980s neural-net paradigm now scores the majority of global card payments — an enormous, uncredited footprint. If it is a garbled marketing figure (e.g. conflating "cards protected" with "transactions scored," or a US-only figure inflated to "the world"), the payoff weakens considerably.
What would answer it
- FICO primary material — FICO annual reports / 10-K filings, or FICO's own Falcon product documentation, stating the share of card transactions its Falcon/neural models score, with a defined denominator (US vs. global; credit vs. debit; accounts vs. transactions).
- An independent audited or analyst figure (Tier 1–2) corroborating the "majority of the world's" framing rather than the softer, more commonly cited "two-thirds of US card transactions" phrasing.
- Resolve the denominator ambiguity: the same lineage is elsewhere quoted as "two-thirds of all card transactions" (the source URL's own title) versus "a majority of the world's" — these are not the same claim.
Until then, the note stays status: seedling with the flag preserved.