Unico responds to ‘DIY Deepfakes,’ fraud spike with authentication orchestration

Unico has put numbers to the dramatic rise in volume of sophisticated and AI-powered fraud attempts this year, and introduced a function to help sort interactions according to the appropriate level of authentication assurance required to increase detection rates without adding friction for low-risk users and transactions.

Smart Revalidation orchestrates authentications between its “silent” method, which does not require user interaction, a passkey and a server-side face biometrics check for high-risk transactions.

The continuous authentication technology is partly motivated by a projected 550 percent surge in AI attacks in 2026, compared to the previous year. That increase led to Unico blocking 23 percent more fraud in the first half of this year than the first half of 2025, as the company revealed within a review of the fraud landscape, which is detailed below.

The new feature identifies patterns and calibrates the appropriate level of authentication by combining over 40 real-time contextual signals. Inputs used by Smart Revalidation include  usage behavior, location, device, biometric and liveness detection history, and transaction history, the company says.

Businesses customize their risk tolerance, validation workflows and compliance rules based on their internal policies.

Unico says its Smart Revalidation approves up to 90 percent of transactions without adding any friction at all, and still performs at 99.9 percent accuracy. Unico says the process is also up to 15 times faster than traditional multi-factor authentication (MFA) techniques.

Every additional step imposed on users can cause abandonment and revenue loss, says Unico GM Fernanda Beato.

“We are entering a model in which identity-applied intelligence becomes continuous,” explains Beato. “We use AI and behavioral signals to determine the required verification level. This allows us to protect highly critical sectors, such as financial services, fintechs, credit, and digital platforms, with greater assertiveness, all without compromising the fluidity of the user journey.”

Sophisticated fraud attempts by unsophisticated attackers surge

Unico blocked 19.6 billion Brazilian reais (approximately US$3.8 billion) in Brazil in the first half of 2026, as the company observed more than 13 million fraud attempts, or 50 per minute.

Professionalized cybercrime and Fraud-as-a-Service (FaaS) models have normalized the use of automated tools and AI in fraud attacks, Beato says, making identity verification “a fundamental strategic pillar for enterprise continuity in the digital economy.”

The plummeting cost of AI fraud was explored in a report released by Unico and Liminal in July.

“DIY deepfakes” have become a significant subcategory of presentation attacks, according to Unico’s data, pushing incidents of relatively simple fraud attacks up by 400 percent in 1H 2026.

Unico’s GM emphasizes the continuous review and stress tests and simulations the company carries out to stay ahead of new attacks.

Article Topics

AI fraud  |  biometric liveness detection  |  biometrics  |  continuous authentication  |  deepfakes  |  Unico