Veriff Launches Presence Verification to Combat Deepfakes and Coerced Fraud
Businesses can now screen identity verification sessions for deepfakes and signs that users are being coached by another person through a new product from identity verification provider Veriff.
Presence Verification combines passive liveness detection with video analysis and device and network signals to identify manipulated media, automated attacks and suspicious activity. An optional coercion detection tier adds audio and video analysis to identify signs of third-party influence.
Users complete verification by recording a short selfie video without submitting an identity document or matching their face against an enrolled image. The system checks for presentation attacks, including photographs, screen replays, and 3D masks, as well as synthetic and injected video.
Device and network analysis helps identify emulators and automated activity. Veriff says its decision engine uses more than 1,000 risk signals.
“Presence Verification goes beyond the face to analyze signals across the device, the network, full video, and audio, to detect injected deepfakes, scripted bots, and spot the signs that someone is being coached,” says Rob Brazier, chief product officer at Veriff.
The additional coercion detection tier analyzes video and audio for signs that another person is directing the applicant. The feature supports 98 languages. Veriff reports more than 98 percent accuracy in detecting coercion and an 84 percent improvement in social engineering fraud detection, though it does not disclose the testing methodology behind the figures.
Businesses can integrate Presence Verification through Veriff’s iOS and Android SDKs, initiating sessions via REST API and receiving automated decisions through webhooks.
Financial institutions can use the technology to detect suspicious account registrations, scam-driven loan applications, and customers being guided through transactions by fraudsters. Marketplaces and dating platforms can introduce checks during registration or before sensitive activities, such as changing payout details or creating high-value listings.
Veriff also sees applications in agentic commerce, where businesses need to confirm human participation during AI-assisted transactions.
The launch follows Veriff’s findings that 4.18 percent of identity verification attempts in 2025 were fraudulent. The Identity Fraud Report 2026 highlighted impersonation accounting for more than 85 percent of detected fraud.
Other fraud prevention providers, including Netarx, Onecom and Pindrop, are also using conversational analysis and real-time monitoring to detect deepfakes and social engineering during customer interactions.