Fraud prevention gets smart(er) with Agentic AI

Agentic AI goes beyond traditional Gen AI by not only generating output but completing actions on behalf of users. This shift is significant enough to redefine e-commerce into A-commerce, where autonomous agents handle end-to-end tasks. In fraud prevention, it represents a new maturity stage, adding reasoning and decision-making capabilities on top of machine learning.

Fraud teams remain cautious about trusting new AI capabilities, especially from vendors. Yet at the same time, organisations are pushing teams to “do more with less,” increasing the pressure to evaluate AI-driven solutions. This dynamic is creating opportunities for AI-native companies to enter conversations with large enterprises earlier than ever before.

Fraudsters are rapidly adopting advanced AI tools that often outperform defense systems. This has caused a surge in automated, scalable attacks like ATO, email takeover and device compromise. The sophistication of these attacks, such as realistic customer journeys and highly convincing document manipulation, creates far more “grey area” for detection.

A practical way to introduce Agentic AI is by breaking work into smaller, clearly defined tasks. Teams can automate pieces such as rule writing, chargeback analysis or data pulls, building confidence while keeping critical decisions human-driven. This approach allows experts to shift their focus to higher-leverage strategic work instead of repetitive analysis.

Traditional bot detection can no longer simply block automated traffic, since many bots now serve legitimate functions. The new challenge is identifying the intent behind each agent interacting with a platform. This is driving the emergence of KYA frameworks focused on verifying, validating and understanding the behavior of both human and non-human agents.

Organisations must develop a clear internal plan for how they will defend against Agentic AI-driven threats and accelerate their decision workflows. External collaboration is equally important, as fraudsters increasingly coordinate efforts more effectively than many defense teams. Companies should challenge their vendors on how they detect and validate agents to ensure readiness for peak-risk periods like the holiday season.