Matheus Raposo, Risk and Performance Consultant at ICTS Protiviti Brazil, explains that the mechanisms involving behavioral biometrics can identify fraud risks before monetization, without interfering in the flow of the user experience on the platform.
Preventing fraud and chargebacks, increasing transaction approval rates and reducing costs with anti-fraud tools are certainly some of the main objectives and key results present in basically all anti-fraud structures of digital platforms, such as e-commerce, fintechs and even digital banks.
However, orchestrating a fraud prevention system that is capable of guaranteeing the overcoming of expectations in all of these aspects can be a challenging task. This is because traditional approaches to transactional anti-fraud tools often operate to make them part of a trade-off.
For example, to ensure a level of chargeback control, approval rates often fall, and this can compromise the user experience and even frustrate loyal customers who are unable to complete their journeys across digital platforms.
Additionally, the effectiveness of the assessments is compromised in transactional anti-fraud approaches, as we have experienced mega data leaks, the largest one being with 223 million Brazilians, that took place in early 2021. As these algorithms basically compare the registration and transactional data of users to the history of transactions previously carried out by the user to guide decision making regarding the approval or disapproval of the transaction, a fraudster, in possession of user data with a good history, can cheat these traditionally used protective layers.
The behavioral biometrics approach assumes that the malicious individual may even be able to defraud the data entered to make transactions but can’t deceive the behavior adopted during the session. Its conduct throughout the entire journey, from the moment you start interacting with the application or the web platform, even before logging in, starts to be evaluated to identify fraud risks before monetization, in other words, when that session becomes a transaction.
This increases reaction time and allows, for example, the creation of challenges to make sure that it is really the account holder making a transaction, and not a fraudster with the aim of financially benefiting from the transactions. All this is without any compromise on the flow of the customer experience on the platform.
Through behavioral biometrics more than 200 user behavior variables are evaluated, from navigation and usage pattern to device and network attributes. All through an Artificial Intelligence algorithm, which in addition to assigning a score for the risk of fraud, offers insights that explain what triggered the alert, allowing the knowledge of fraud cases and modus operandi, and also facilitating the integration with other solutions of this kind.
Another point worth mentioning is that usually the cost of transactional anti-fraud tools varies and depends on the number of transactions analyzed in the company’s environment, whether they are approved or not.
Thus, by adopting an additional layer of protection that analyzes session characteristics, it is possible to reduce the volume of analysis requested for transactional anti-fraud solutions, reducing total costs, since the license for the behavioral biometrics tool is annual and has a fixed cost agreed at the beginning of the period, with no surprises or additional fees.
It is recommended initially to design a cost savings based on those higher risk score sessions. In a second stage, with greater security in the behavioral biometrics layer and with the advance of algorithm training to identify transactions with low fraud risk, it is common for companies to also give up the analysis of transactional tools for those sessions with a risk of very low fraud, further optimizing the cost structure with anti-fraud.
Even for those sessions with an average fraud score where the most recommended approach continues to be the use of at least one more layer of protection composed of a transactional anti-fraud tool, there are solutions whose proposal is to guarantee the highest possible pass rate, committing to the refund of all chargeback due to the approval of false positives.
Therefore, through the orchestration of a fraud prevention system composed of integrated layers of protection considering the entire journey, it is possible to overcome the challenge of increasing the approval rate, controlling fraud and chargebacks and also reducing costs with transactional anti-fraud tools without compromising customer experience on digital platforms.
