Lending without fraud, without friction.

Synthetic identity fraud combines a real Social Security Number with a fabricated name, address or DOB. We attack it on three fronts: device and behavioral signals at session start, document verification with OCR and template-tampering detection, and a 1:N biometric search that flags whether the same selfie has already been used under a different identity. We also screen the SSN against velocity rules and our biometric blacklist, which is the single biggest predictor of synthetic-identity loan defaults.

Yes. Regulation B (ECOA) prohibits discrimination based on race, color, religion, national origin, sex, marital status, age or receipt of public assistance. KYC data points such as photo, name and address can correlate with protected classes, so you must isolate the underwriting model from those features and document that the model does not use them as predictors. Legal Talent's KYC output is structured so credit-decisioning teams can ingest the verification result without bleeding through demographic features.

The FFIEC Authentication Guidance acknowledges biometrics as a strong authenticator and the BSA CIP rules allow non-documentary verification methods that produce a reasonable belief of identity. Most US digital lenders combine government ID capture with a liveness-checked selfie and a 1:1 face match: that combination is well-accepted by examiners and by warehouse-line counterparties. Legal Talent's NIST-aligned liveness and face-match models are calibrated for the FAR/FRR thresholds typically required by lending warehouses.

Every applicant is screened against OFAC SDN, OFAC Consolidated and OFAC's sectoral programs at application time and again at funding. We use fuzzy matching tuned for nicknames and transliteration and we re-screen the entire borrower book daily against new designations. Hits are routed to a review queue with the supporting evidence (alias, DOB, geography), and the lender keeps full control over auto-block versus manual-review thresholds.

Yes. Our automation rules engine runs at the end of the KYC session and emits a verdict (approve, decline, escalate) within two seconds. You can mix signals from document verification, liveness score, sanctions outcome, adverse-media classification and your own credit-bureau pull. Lenders typically auto-approve about 75 percent of clean applicants and escalate the rest to a human reviewer with the full evidence pack pre-loaded.