KYCtruth
A real-time deepfake shield for video KYC — challenger banks and fintechs get an instant trust score on the customer on the other side of the video call.
compliance officers at UK and EU challenger banks running video KYC onboarding
- Browser plug-in that scores each onboarding video stream for synthetic artefacts in real time
- Side-by-side liveness check against the selfie ID uploaded seconds earlier
- Auto-logs flagged sessions with timestamps and PDF evidence pack ready for SAR filing
- Admin dashboard tracking flag rate by region, device, and channel so fraud teams can tune thresholds
AI-proxied applicants are breaking video KYC for fintechs, and off-the-shelf detectors have not been packaged for compliance teams.
Regula 2024 survey found 49% of businesses have encountered audio/video deepfake fraud; UK payment fraud losses hit £1.28B in 2025 with banks warning of AI-generated deepfakes; Banker and Forbes both flag deepfakes actively disrupting KYC in fintech.How Deepfakes Are Disrupting KYC And Financial Security ↗AI-driven fraud costs UK £1.28 billion as banks warn of deepfake surge ↗
Space is already crowded: FrameSentinel is explicitly 'purpose-built for video KYC fraud detection' with 5 detection modules and sub-2s API; Sumsub, Persona, Onfido, FaceTec, Incode, Jumio, Veriff, and InsightFace all ship deepfake/liveness defenses; FrameSentinel appears to have already taken the exact 'compliance-team packaged detector' positioning the idea claims is unserved.Best Video KYC Fraud Detection Tools 2025 | FrameSentinel vs Persona vs Onfido ↗Top 4 AI Fraud Prevention Platforms for Crypto Exchanges in 2026 ↗
Per-check pricing is established (Sumsub from $0.10/check, deepfake APIs $99/mo Pro with enterprise custom); challenger banks have real compliance budgets and FCA-driven pain (Monzo £21m compliance issue), but WTP for a standalone 'trust score' on top of existing Onfido/Sumsub contracts is unproven and competes with already-embedded vendors.Pricing | Deepfake Detection API ↗When QA breaks down: inside Monzo's £21m compliance ordeal ↗
AI-proxied fraud is a structural, escalating threat (Forbes/Clifford Chance/Regula all treat it as a multi-year horizon) and FCA scrutiny of challenger-bank financial-crime controls is permanent; however, liveness-first platforms and upcoming regulation could shift the 'detector layer' back into bundled IDV suites over 3-5 years.Financial services face up to deepfake risks - Clifford Chance ↗Financial crime controls at challenger banks | FCA ↗
Underlying deepfake models (InsightFace, open detectors) exist, but real-time sub-2s inference on a live video stream, false-positive tuning for legitimate customers, and integration into live agent video KYC stacks is genuinely hard engineering; FrameSentinel claims sub-2s/99.7% but a new entrant still needs significant ML and low-latency infra work.Enterprise Deepfake Detection SDK & API | InsightFace ↗Best Video KYC Fraud Detection Tools 2025 ↗