Financial fraud is a multi-billion-dollar problem that is only getting more complicated for digital-native organizations like EVO Banco to solve.
As one of the key challengers in Spain’s financial services industry, EVO Banco is known for disrupting the national banking sector with its 100% digital banking model. The company’s goal? Using innovative technology to create modern banking solutions that make customers’ lives simpler.
To fulfill that mission, EVO Banco needed a cloud-native solution for data streaming. Find out why the bank chose Confluent Cloud for its data infrastructure and how the bank uses the platform to manage large volumes of data in real time, delivering a better customer experience.
As EVO Banco’s Chief Data Officer and Manager of Innovation, Jose Enrique Perez is well-versed in how the company uses machine learning to both prevent fraud and improve the customer experience in real time.
Jose and Confluent’s Duncan Ash, Vice President of Global Industries, recently sat down to discuss how EVO Banco is leveraging real-time solutions to stay one step ahead of sophisticated fraudsters.
During this online talk, Jose explained some of the most common methods of financial fraud, the unique challenges facing digital-native banks, and how EVO Banco is using real-time data streams to power advanced fraud detection. This blog post recaps lessons learned from Jose’s talk—read on for details and tips on using data streaming to fight sophisticated fraudsters with sophisticated technology.
Today’s financial institutions are inundated with evolving and emerging security threats. To protect their brand reputation—and their customers—banks need to develop comprehensive fraud detection and prevention systems.
It’s not enough to just prevent an active attack; banks need solutions that allow them to preempt the next attack. Digital banking services in particular are a primary target for bad actors, which means they need to be proactive in how they identify and address fraud—while not compromising the customer experience.
Bank fraud is a crime that takes many forms, and to protect their customers and their systems, financial services organizations need to stay aware of the most common methods.
Often, the most rampant fraud methods are the simplest. For example, phishing is a common technique used to trick individuals into revealing sensitive information, such as login credentials or financial data. Phishing for this information can happen over email, over a phone call, or even text message.
This variety of potential attack vectors makes this seemingly simple method of financial fraud difficult to prevent using fraud detection technology. Other common methods of fraud include identity theft, account takeover, and credit card fraud. To prevent these crimes, banks rely on sophisticated authentication methods and monitor accounts for suspicious activity.
When able to process and react to data in real time, advanced fraud detection systems can identify unusual activity and flag potentially fraudulent transactions and account usage. Banks can also use these systems to identify suspicious activity that indicates a customer has been the victim of an investment scam.
Because there are so many different types of bank fraud, banks need to implement advanced fraud detection systems that combine real-time monitoring, advanced authentication methods, and real-time analytics.
Additionally, protecting customers in a fully digital environment is often difficult because banks have to do the following, all without in-person interactions:
Authenticate customers and provide secure access to their accounts
Rapidly identify and handle fake or compromised accounts
Stop and prevent payment fraud
Detect and protect their customers from scams
Whether identifying suspicious account activity, credit card transactions, or digital payments, today’s banks are ingesting and analyzing large volumes of continually generated data. This is why artificial intelligence and machine learning (AI/ML) models have become integral parts of modern fraud prevention systems.
To protect its customers, EVO Banco uses a variety of big data tools to identify anomalies in transaction data and account activity, including ML algorithms and rules-based systems. When the bank’s fraud detection system identifies suspicious activity, real-time alerts notify end users.
These real-time alerts—sent via email, SMS, or mobile push notifications—allow customers to act quickly and respond to fraudulent activity. This immediacy not only safeguards the customers’ accounts but also protects EVO Banco’s reputation and trust.
EVO Banco’s fraud detection system relies on two main technologies: Confluent Cloud for real-time transaction processing and Apache Spark for batch processing and ML model training. The bank uses Spark to process historical data in batches to train the fraud detection system’s ML model.
In real time, Confluent Cloud receives and processes transaction data from multiple sources, including ATMs, online payments, and mobile banking. The data streaming platform immediately sends that data for real-time processing so transactions can be analyzed for fraudulent activity using the ML model trained with historical data.
This mind map shows how the fraud detection system sends data from various sources, including transaction records, customer information, and account details. This data is all streamed through Confluent Cloud so it can be used to train the system’s ML models and trigger real-time alerts when fraudulent activity is detected.
EVO Banco’s advanced fraud detection system relies on Confluent Cloud to analyze large volumes of data in real time and put those insights into action. This system allows the bank to combine behavior analysis, data analytics, and predictive modeling to protect against financial crime—without interrupting or sacrificing the customer experience.
Protecting the quality of that customer experience is incredibly important to EVO Banco, as the organization takes pride in providing seamless, reliable service to its users.
Choosing Confluent Cloud to power the organization’s data infrastructure allowed EVO Banco to:
Maintain robust data quality management and governance policies
Aggregate a variety of data sources across multiple systems (e.g., transaction data, customer data, device data, authentication logs)
Provide contextual, high-fidelity data to feed AI/ML models
Process both real-time and historical transaction data, routed through a single data streaming platform
Further, Confluent Cloud has allowed the bank to implement in-stream fraud detection for more than 500, 000 daily transactions. That’s allowed EVO Banco to significantly increase the accuracy of its fraud algorithms and reduced reaction times to just seconds.
Since adopting Confluent Cloud, the bank has reached an average of 500 fraudulent transactions blocked daily and reduced weekly fraud losses by 99%.
Ready to learn more about how data streaming is transforming financial services? Then, you’ll want to attend Current 2023: The Next Generation of Kafka Summit, September 26-27, 2023, in San Jose, California.
This two-day event will give you plenty of opportunities to dive into the architectures, technologies, and use cases transforming your industry. At Current 2023, you’ll get to hear from experts who are developing innovative financial services solutions with data streaming.