Deepfake detection for remote customer onboarding

Identify injected video, face swaps and synthetic faces in real time, inside your own data centre. A second layer next to the liveness check you already run.

On-premisesReal timeExplainableREST API
Why now

The attacks that get through today bypass the camera

1 in 5

biometric fraud attempts now use a deepfake

Entrust 2026 Identity Fraud Report

+40%

injection attacks year on year

Entrust 2026

+741%

injection attacks on iOS in 2025

iProov Threat Intelligence 2026

100,000+

injection attempts per month against onboarding flows in Africa

Smile ID 2026

Liveness checks were built for photos, replays and masks held in front of a camera. Today's attacks replace the camera stream itself, or put a swapped face on a live operator who answers every challenge. Regulators across Türkiye, the Gulf, Central Asia and Africa are responding by mandating liveness and local processing, and moving fraud liability onto banks.

Sources: Entrust 2026 Identity Fraud Report (Nov 2025); iProov Threat Intelligence Report 2026 (Apr 2026); Smile ID Digital Identity Fraud in Africa 2026 (Mar 2026).

Product

One verdict per session, from inside your perimeter

Your KYC engine or agent application sends a short video clip; hatkoai. returns a verdict, a score and a plain-language reason in under a second. Nothing leaves your network.

Your application

KYC engine, agent desktop or mobile app sends a short clip

hatkoai. engine

Deepfake checks, on your servers

Analysis result

The result and score are shown in your rules engine or on the agent's screen

Integration through a REST API. Works with the KYC and video-call platforms you already have.

Coverage

Detection in every channel where a face is verified

1

Remote customer onboarding

Video-call onboarding with a live agent, and automated onboarding flows. The agent sees the analysis result and score next to the customer's video, in real time. As AI agents begin to run these calls, the same layer protects them too.

2

Face recognition and KYC flows

Wherever face recognition or a KYC step is triggered (login, password reset, new-device binding, account recovery, and high-value approvals such as loans, limit increases and large transfers), a safer verification with a deepfake check added.

Who it is for

For every system that onboards or grants access with face recognition

Banks
Fintech and payment institutions
Identity-verification platforms
Video-call platforms
Telecom operators
Public e-ID services

Why hatkoai.

1

Real time

Live calls are analysed in real time, so a transaction can be approved or rejected on the spot.

2

Resilient

Deepfake generation algorithms evolve constantly; regular software releases keep the system resilient against them.

3

Data stays in-house

Installed inside the institution; an auditable structure that is ready for inspections.

For your engineering team

Easy integration, on-premises deployment

hatkoai. is installed on the institution's own servers. Your KYC application connects to the hatkoai. API for verification; no additional infrastructure is required. Your team receives all documentation needed for installation and integration. Every deployment begins with a pilot in the institution's own environment.

Talk to us about a pilot
Contact

Test it on your own onboarding videos

Send us a message and we will set up a pilot in your own environment.

Address

Dijitalpark Teknokent, Kirazlıdere Mah. Eski Ankara Cad. A-1 Blok No:4A İç Kapı No:19, Çekmeköy / Istanbul, Türkiye