# Fake Payslip Detection for Loans > Income is the primary underwriting variable, and payslips are easy to fake. TamperCheck catches the three patterns that dominate loan fraud: net pay inflation on a genuine payslip, fully fabricated payslips from templates, and AI-generated payslips from text prompts. All three pass visual review. Forensic analysis catches what arithmetic and eyeballs can't. **Canonical URL:** https://tampercheck.ai/fake-payslip-detection-for-loans **Last updated:** 2026-10-05 --- ## Why payslips are the go-to fraud target in lending **Payslips are the most-edited document in loan applications.** Income is the primary underwriting variable, and a payslip is easy to fake: most look similar, the structure is well known, and the figures are small enough to feel low-stakes to fraudsters. The result is that payslip fraud accounts for the largest share of document fraud across personal lending, mortgages, and auto finance. Three patterns dominate: **net pay inflation** (changing a single number), **fabricated payslips** (generated entirely from a template), and **AI-generated payslips** (created by a generative model from a text prompt). All three pass visual review. Even bank-statement cross-checks are unreliable - fraudsters know to inflate both documents consistently. Fake payslip detection has to be forensic, not arithmetic. The signals are in the PDF structure, the font metrics, the producer metadata, and the spectral properties of any embedded images. TamperCheck inspects all of them automatically. ## What TamperCheck detects ### Net-pay arithmetic verification Gross pay, deductions, tax, super/401k, and net pay must all reconcile. A single edited figure breaks the chain; TamperCheck catches it. ### Font metric consistency Edited fields often use slightly different font weights or spacing. Cross-field font metric comparison reveals these edits. ### Producer metadata audit Real payroll systems (ADP, Xero, Gusto, MYOB, BambooHR) produce PDFs with consistent producer signatures. Edits introduce new tool fingerprints. ### AI-generated payslip detection Fully synthetic payslips from generative AI tools have spectral signatures that real payroll-system PDFs don't. The CV layer is trained on these. ### Year-to-date plausibility YTD figures, pay period sequencing, and date consistency are cross-checked against the claimed pay frequency. ### Bank statement cross-validation (optional) When the borrower submits a bank statement, TamperCheck cross-validates that the net pay actually credits the account on the claimed date. ## Payslip fraud detection that sits inside your loan origination workflow Most lenders wire it into document collection. Every payslip an applicant uploads goes through the API. The verdict gates the application's progress to underwriting. - **Integrates with any LOS** - Encompass, Calyx, nCino, MeridianLink, custom - **Runs on every payslip** - not just flagged applications - **$0.50 per document** - a fraction of any approved fraudulent loan - **Verdict in about a minute** - keeps origination workflows moving - **Risk score for auto-routing** - clean payslips skip manual review - **Works with multiple payslips per applicant** - cross-validates across them - **Zero storage** - critical for borrower data compliance ## Frequently asked questions ### How do you detect a fake payslip on a loan application? Real detection requires forensic analysis, not just arithmetic. PDF structure, metadata, font metrics across every field, pixel-level signals on embedded images, the producer signature from the payroll system, and AI-generation spectral signatures. TamperCheck runs all of these automatically. ### What are the most common signs of a fake payslip? Producer metadata that doesn't match the claimed employer's payroll system, font metric inconsistencies between fields, net pay that doesn't reconcile with gross minus deductions, AI-generation noise patterns, and YTD figures that don't align with the pay-period sequence. ### Can you detect AI-generated payslips? Fully synthetic payslips from generative AI tools carry spectral signatures and noise patterns that real payroll-system PDFs don't have. TamperCheck's CV layer catches these even when the payslip passes visual inspection. ### How much does payslip fraud detection cost per loan? $0.50 per payslip. Most loan applications include 1-3 payslips, so $0.50-$1.50 per application. A fraction of any single fraudulent loan that would otherwise be approved. ### Does payslip fraud detection slow down loan origination? Verdicts return fast enough that checks complete while the applicant is still on the next application step. Clean payslips can be auto-approved at the document level, so manual review only sees the suspicious cases. ### Can it work with payslips from any country? The forensic checks are document-agnostic - PDF structure, font metrics, producer metadata, and AI-generation signatures apply equally everywhere. Country-specific tax-line validation kicks in automatically when the document is classified. ## Related use cases - https://tampercheck.ai/use-cases/lending-credit - https://tampercheck.ai/use-cases/tenant-screening - https://tampercheck.ai/use-cases/hr-hiring ## Compare with alternatives - https://tampercheck.ai/compare/inscribe-vs-tampercheck-ai - https://tampercheck.ai/compare/ocrolus-vs-tampercheck-ai - https://tampercheck.ai/compare/snappt-vs-tampercheck-ai --- Try TamperCheck now: https://tampercheck.ai/dashboard · Docs: https://tampercheck.ai/docs · Pricing: https://tampercheck.ai/pricing