Detectory
All industries

Financial services

Document fraud detection for KYC and loan underwriting

KYC onboarding and underwriting run on the documents fraudsters fake best: paystubs, bank statements, and IDs. Detectory renders a forensic verdict on every uploaded document and resolves the synthetic identities behind them, before the account opens or the loan funds.

The problem

Synthetic identity is the fastest-growing financial crime in the US

Synthetic identity fraud drives an estimated $6B+ in annual lender losses, and generative tools have made fake paystubs and bank statements a commodity: template farms sell "verified income packages" for under $50. Manual document review in onboarding and underwriting can’t keep pace, and every fake that clears KYC becomes a charge-off, a SAR filing, and a regulator conversation later.

$6B+

annual synthetic-identity lender losses

Federal Reserve estimate

<$50

street price of a fake income package

fraud marketplace monitoring

85%

of synthetics pass traditional KYC checks

industry studies

Example verdicts

The documents your team sees, scored

Representative documents from this workflow with the forensic verdict Detectory renders and the finding behind it. Suspect documents are never auto-denied; they route to your reviewers with the evidence itemized.

Bank statement (underwriting)

Three months of statements supporting a loan application

Likely fraudulent

Finding: Running balance does not reconcile across pages; font substitution on deposit lines; PDF assembled from two source documents

Paystub (income verification)

Gusto-format stub in a personal-loan file

Likely fraudulent

Finding: Template matches a known stub-generator service; employer domain registered 11 days before application

Driver license (KYC)

Government ID for account onboarding

Genuine

Finding: Barcode, MRZ-equivalent data, and security-feature geometry consistent; cleared in 6 seconds

W-2 (mortgage file)

Prior-year W-2 supporting stated income

Human review

Finding: Arithmetic valid but employer EIN newly issued and address is a virtual office; routed to underwriter with findings

What Detectory does here

Purpose-built detection for this workflow

Onboarding document verdicts

Genuine-or-fraudulent calls on IDs, paystubs, and statements in seconds, inline with your KYC flow, so good customers are not slowed down.

Synthetic-identity detection

Identity-fragment analysis catches synthetics assembled from real SSNs and invented biographical data that pass bureau-based checks.

Fraud-ring clustering

Shared templates, devices, and employer artifacts across applications expose organized first-party and bust-out rings.

SAR-ready evidence trails

Every verdict carries an itemized, timestamped evidence chain your BSA/AML team can lift directly into SAR narratives and exams.

Case walkthrough

A bust-out ring warms up 40 accounts

Over three months, a ring opens checking accounts and small credit lines using synthetic identities backed by generated paystubs. Each application passes bureau checks. Detectory flags the shared stub template and a common employer EIN across 40 files, clusters the device fingerprints, and scores every new application from the ring at submission.

Outcome: The fraud team freezes the cluster before the coordinated bust-out, and the BSA team files SARs with the full documentary evidence chain attached.

Run your last quarter’s confirmed-fraud files through Detectory and see what it catches.

Other industries