TrueDoc: Document Fraud Detection for the Generative Era
AI made convincing fake paystubs free to produce. TrueDoc is the countermeasure: ensemble detection for forged and AI-generated documents at intake.
TrueDoc detects forged, tampered, and AI-generated documents: paystubs, bank statements, IDs, invoices, tax forms. It runs a multi-model ensemble where independent detectors cross-validate findings, flags synthetic documents produced by image and text generators, works with documents from any country, and deletes originals after analysis, keeping only redacted reports. Freemium to start; the upper business tiers expose a REST API with batch processing and webhook callbacks.
The enterprise angle is intake. Every lender, insurer, landlord platform, and payroll verifier runs on documents submitted by strangers, and generative AI just dropped the cost of a convincing fake paystub to zero. Manual review was already losing that race; now it is not even in it. Scored through the COST/WORK/FACTORY lens this is a cost and work story: it automates a verification step that either did not happen or consumed analyst hours, in exactly the document classes where fraud is industrializing.
The adoption caveat is false positives. A wrongly rejected legitimate applicant is a regulatory and reputational problem, not just a support ticket. First step: run it in shadow mode against a month of real intake documents, measure the flag rate and where flags cluster, and design the human appeal path before wiring any automatic rejection.
Written by Adib Kadir. Product and engineering executive focused on rolling out AI at enterprise scale.
Start a conversation