Sumsub launches first AI Agent Verification to combat rising deepfake fraud across Africa

Global verification leader Sumsub has unveiled its pioneering AI Agent Verification solution, introducing a Know Your Agent (KYA) framework to combat AI-enabled fraud.

RELATED: Exabeam launches AI security workflows to govern AI agents & mitigate risk

The solution links automated AI agents to verified human identities, creating accountability as artificial intelligence reshapes digital platforms across Africa’s fintech, e-commerce, and service sectors.

Addressing Africa’s “Sophistication Shift” in Fraud

The launch responds to a critical trend identified in Sumsub’s Identity Fraud Report 2025–2026. Africa is experiencing a “Sophistication Shift,” with a 180% year-on-year rise in coordinated, multi-step fraud attacks.

While overall fraud rates declined in mature markets like Kenya, deepfake fraud attempts surged dramatically—up 367% in the DRC, 317% in Tanzania, and 269% in South Africa—indicating fraudsters are deploying more advanced, AI-assisted techniques.

From Blocking Automation to Managing Risk

Many platforms currently block all automation to mitigate risk, creating user friction. Sumsub’s solution takes a different approach by enabling risk-based control. It verifies the human behind an AI agent, allowing trusted automation to function while isolating malicious activity.

ADVERTISEMENT

“Rather than attempting to blindly trust AI agents themselves, our solution focuses on verifying the humans behind them,” said Vyacheslav Zholudev, Co-founder and CTO at Sumsub.

How AI Agent Verification Works: A Four-Pillar Framework

The system builds on Sumsub’s full-cycle verification platform to provide layered security:

1. Device & Bot Detection: Identifies automated activity and assesses real-time risk.
2. Mule Network Prevention: Uncovers coordinated fraud patterns across devices and accounts.
3. Liveness Verification: Confirms a real human is present at critical moments (e.g., onboarding, high-value transactions).
4. Continuous Risk Scoring: Monitors behavior to allow, limit, or challenge automation based on dynamic risk levels.

This model ensures automation never operates without human accountability, protecting platforms from scalable AI-driven attacks while reducing friction for legitimate users.

Enabling Secure Innovation in Digital Ecosystems

“For African businesses scaling rapidly, automation is not optional,” said Artem Popov, Head of Fraud Prevention at Sumsub.

The solution allows organizations to adopt AI-driven efficiency without enabling anonymous abuse, securing the backbone of future digital operations across the continent.