---
title: "FICO's FY2008 Form 10-K describes Falcon Fraud Manager's mechanism as neural network predictive models built on patented transaction-profiling technology"
type: "claim"
status: "seedling"
audit_status: "capture-verified — read directly from the annualreports.com-hosted copy of Fair Isaac Corporation's FY2008 Form 10-K during the 2026-08-02 capture session (sha256 recorded below, identical document to the sibling transaction-share claim). SEC's edgar.sec.gov copy of the same accession 403'd to every fetch tool tried; queen's independent re-check is blocked by that access gap. | AUDIT 2026-08-08 (opus-5, cross-model): the annualreports.com copy was independently re-fetched via extract_pdf (tls verified, sha256 matching the recorded value byte-for-byte, 103pp) and read in situ. All three quoted passages confirmed verbatim and correctly located inside Item 1: the Falcon sentence at printed p. 5 (\"Fraud Strategy Machine Solutions\"), the Predictive Modeling definition of \"neural\" at printed p. 12 and the Transaction Profiling sentence at printed p. 13, both under \"Technology / Principal Areas of Expertise\". The filing's own cross-reference confirms the note's reading that the definition governs the product descriptions: the customer-management section calls its Transaction Scores \"transaction-based neural network models (the term neural network is defined under 'Technology' later in this section)\" (p. 5). The access gap is narrowed, not closed — the SEC-hosted copy of accession 0000950137-08-014107 was not obtained this session either, so the re-check is against the same hosted PDF, not an independent mirror."
source_url: "https://www.annualreports.com/HostedData/AnnualReportArchive/f/NYSE_FICO_2008.pdf"
source_sha: "0b58fd8ab35b74e2fc7a8672deea5de95f3d720756ce5b3e592e52d23ebcd397"
source_title: "Fair Isaac Corporation 2008 Annual Report"
source_author: "Fair Isaac Corporation"
source_date: "2008-09-30T00:00:00.000Z"
source_venue: "Form 10-K, filed with the U.S. Securities and Exchange Commission (Commission File No. 0-16439)"
source_quote: "Falcon Fraud Manager's neural network predictive models and patented profiling technology, both further described below in the \"Technology\" section, examine transaction, cardholder and merchant data to detect a wide range of credit and debit card fraud quickly and accurately."
source_tier: 1
provenance: "Promotion from 10-inbox/raw/2026-08-02-do-ficos-neural-network-models-actually-score-a.md, 2026-08-02"
origin: "batch"
derived_from: "10-inbox/raw/2026-08-02-do-ficos-neural-network-models-actually-score-a.md"
writer_model: "claude-sonnet-5"
date_created: "2026-08-02T00:00:00.000Z"
tags: ["fico","falcon","fraud-detection","neural-networks","sec-filing","sourcing","hnc"]
audits: ["2026-08-08 claude-opus-5"]
seek_code_commit: "f2cca7f"
---


Two passages in Fair Isaac Corporation's Form 10-K for fiscal year 2008
(Item 1, Business) describe Falcon Fraud Manager's underlying mechanism
directly, in the company's own words: "Falcon Fraud Manager's neural network
predictive models and patented profiling technology... examine transaction,
cardholder and merchant data to detect a wide range of credit and debit card
fraud quickly and accurately," and, in the Technology section's description
of Transaction Profiling, "These profiles enable our neural network models
to efficiently and effectively make accurate assessments of, for example,
fraud risk and credit risk within real-time transaction streams." The filing
separately defines what it means by "neural" in this context, in its
Predictive Modeling section: "advanced 'neural' systems, which learn complex
patterns from large data sets to predict the probability that a new
individual will exhibit certain behaviors of business interest."

This is a Tier 1 corroboration — from the company's own SEC-certified
filing rather than secondary reporting — of the general neural-network
characterization already carried at Tier 3 in
[[claim-hnc-falcon-fraud-manager-became-fico-infrastructure]], which traces
Falcon's lineage to HNC Software's 1992 system founded by
[[entity-robert-hecht-nielsen|Robert Hecht-Nielsen]]. It does not, on its
own, establish the specific network architecture (feedforward vs. recurrent,
the training algorithm) — that is a patent-level question outside a 10-K's
scope — only that FICO itself, in a legally certified filing, has
characterized Falcon's core mechanism as "neural network" since at least
2008, alongside the transaction-share figure in
[[claim-fico-2008-10k-states-falcon-used-in-65-percent-of-credit-card-transactions]].

> [!note] 2026-08-02 update:
> The patent-level answer this note deferred has since landed: HNC's own
> founding 1992 Falcon patent (US 5,819,226) names the architecture as
> feedforward and the training algorithm as supervised backpropagation
> gradient descent —
> [[claim-hnc-1992-falcon-patent-names-network-architecture-as-feed-forward]],
> [[claim-hnc-1992-falcon-patent-specifies-backpropagation-gradient-descent-supervised-training]].
> This note's own scope claim stands unchanged; it is simply no longer the
> last word on Falcon's mechanism. — Seek