BioCatch Launches AiRE to Detect Agentic AI Fraud Risk

What happened

BioCatch, a company that prevents fraud, scams and financial crime by recognising patterns in human behaviour, has launched BioCatch AiRE™ (Agentic AI Risk Essentials), its new agentic AI detection capability. Finextra reported the launch.

AiRE extends BioCatch's focus into agentic AI, the category of AI systems that can carry out multi-step tasks with limited human direction. The reporting available to us does not set out technical specifics, availability, pricing or named customers, so those details should be treated as unconfirmed until BioCatch publishes them.

Why it matters

The launch signals that risk teams are being asked to detect not only suspicious people but also suspicious software acting on their behalf. When automated agents can initiate payments, open accounts or interact with service channels at speed, the signals that separate a legitimate customer from a fraudulent one become harder to read from transactions alone. This points to a shift from checking what is being done towards assessing how it is being done, and by whom or what.

For leaders in experience and digital transformation, that shift has direct consequences. Every new layer of detection sits inside a customer journey, and the way it is designed determines whether protection feels like care or like obstruction. Fraud controls that operate invisibly can preserve trust; those that interrupt users without clear reason tend to erode it.

The Renascence take

Most readers will focus on the fact that an AI product is being used to catch AI. The more important question is what customers feel when a defence engages, and whether the friction it creates is proportionate to the risk. Good behavioural design treats intervention as a scarce resource, reserved for moments where the signal genuinely justifies it. A customer-obsessed operator should map where agentic activity meets its own journeys, identify which moments would be most damaging to get wrong, and test how each control feels from the user's side before deciding it is working.

Agentic AI does not just raise the stakes for fraud teams; it exposes how little most organisations have designed their controls around experience. The real test of AiRE, or any tool like it, will be whether legitimate customers notice it at all. Protection that is felt as a dead end has already failed, however accurate its detection. Operators should treat every new detection layer as a service design decision, not only a security one.