Digital businesses are under significant pressure to enhance fraud prevention measures in the face of increasing AI-assisted fraudulent activities. The rise of synthetic identities and automated document generation challenges traditional verification methods, enabling organized groups to exploit system vulnerabilities swiftly.
Know Your Customer (KYC) practices are particularly at risk, with skilled fraudsters leveraging AI to imitate legitimate customers. "Currently, KYC is the most targeted layer, where real-time document generation and deepfake face-swapping are becoming remarkably sophisticated," stated Volodymyr Todurov, CEO of Frogo, a provider of fraud prevention and risk management solutions.
However, implementing strict controls can have detrimental effects on businesses. While these measures may help prevent fraudulent transactions, they can also lead to the rejection of genuine payments, create delays in withdrawals, and ultimately push customers to seek alternatives. As a result, companies might report lower fraud rates yet lose valuable customers in the process.
Frogo combines device intelligence, automated risk scoring, and real-time alerts, advocating for a reassessment of how businesses approach fraud prevention. "Operating with a 'guilty until proven innocent' mindset while relying on aggressive, automatic blocks can completely destroy the customer experience," Todurov warned.
This perspective urges a deeper look at anti-fraud effectiveness by considering profitability alongside prevented losses, operational costs, and the impact on legitimate clients.
Todurov refers to anti-fraud strategies as a "revenue optimisation engine," evaluated through three interconnected metrics: direct loss prevention, operational efficiency, and retention and lifetime value. Direct loss prevention, the most recognized metric, tracks the financial losses attributed to fraud, including illegitimate withdrawals, and assesses whether those losses are decreasing. However, it is essential to consider the resources required to achieve that reduction. A business faced with increasing customer volumes may find it challenging to function efficiently without expanding its risk management team.
"Without a nuanced evaluation model, you risk blocking your highest-revenue players," he cautioned.
Operational efficiency can also be measured by the number of monthly active users each analyst supports and the time invested in completing investigations. Automating routine decision-making processes should enable specialists to focus on addressing more complex threats.
The third metric evaluates effects on legitimate customers. Todurov supports policies for "automated withdrawals based on the presumption of innocence," suggesting that smoother transaction processes can foster trust and enhance long-term customer value. Delays or unnecessary checks can incur costs, even if they don't show up in fraud reports.
These three measures should be evaluated collectively. A system may effectively flag some fraudulent activities while overlooking significant fraud elsewhere or implement controls that alienate genuine patrons. The financial implications hinge on recognizing what the system overlooks and the costs associated with its interventions.
High-value customers exemplify the potential pitfalls of aggressive fraud controls. Their behaviors may sharply differ from average users due to frequent travel, leading to sudden location changes, or large deposits that can trigger spikes in activity. Such signals could easily be mistaken for fraudulent actions, complicating the business's efforts to retain its most profitable clientele.
"Without a nuanced evaluation model, you risk blocking your highest-revenue players," Todurov reiterated.
Frogo advocates for a more informed approach to atypical behaviors. Depending on the situation, certain signals could necessitate intervention by support or retention teams rather than immediate rejection.
Distinguishing these nuances requires specialists with a thorough understanding of both the product and its customer base. Their expertise helps determine where AI can enhance decision-making processes and where traditional methods remain more effective. "AI isn't a silver bullet," remarked Todurov. "Its effectiveness relies on specialists who understand where machine learning can add value and where time-tested, deterministic logic still prevails."
To make informed decisions, specialists need comprehensive information from a user's initial visit, beginning with the device they utilize to access the platform. Frogo emphasizes the importance of device fingerprinting, analyzing both hardware and software parameters to authenticate user legitimacy, thereby identifying bots, spoofed devices, or masked origins.
Frogo offers a standalone Device Fingerprinting tool that allows businesses to incorporate this feature into their existing anti-fraud measures. Businesses with basic in-house rules but lacking device recognition can benefit significantly from this application. Device context assists analysts in determining if a sudden change in location signifies a returning customer or warrants further examination.
"Anti-fraud tools should be seen as core data engines," said Todurov. The tool evaluates over 300 client-side identifiers, helping maintain continuity within a single domain and recognize devices across multiple brands.
This tool also contributes contextual data, including IP risk, VPN and proxy usage, browser specifications, screen resolution, and hardware characteristics. Other signals like time zone discrepancies, remote desktop software, and virtual machine detection are also taken into account.
Within Frogo's broader platform, actions such as registration or transaction events are cross-checked against the business's risk criteria. Todurov mentioned that the system reacts in under a second, detailing which risk triggers were activated.
The value of this data collection extends beyond mere threat identification; it also aids businesses in recognizing and leveraging positive customer behavior. "Anti-fraud tools should be viewed as core data engines. Instead of only scoring bad behavior, you can define triggers that identify high-value, 'ideal' player behavior," he stated.
Frogo’s real-time alerting system directs negative risk triggers to analysts while transmitting positive behavioral signals to marketing and VIP teams via webhooks or Telegram. This allows various departments to respond to underlying activities aligned with their specific responsibilities.
As businesses venture into new markets, they must continually adapt their strategies. Customer behaviors, payment preferences, and fraud tactics may differ, necessitating revisions to previously effective rules.
The primary aim is to mitigate true threats while minimizing unnecessary workloads and ensuring smooth experiences for legitimate customers as the business grows. "The winners will combine a highly skilled fraud team with agile technology, viewing anti-fraud through a growth mindset," Todurov concluded.
