How Wipro Improves Financial Fraud Detection with Real-Time Fraud Detection Using IFFD

By Venkatesh Balasubramaniam, Financial Crime Compliance Specialist – Wipro
By Prajeesh Kozhisseri, Head of AWS Engineering Solutions – Wipro
By Stephen Randolph, Senior Partner Solutions Architect – AWS

Wipro

The rate of financial fraud & scams continues to increase globally, posing significant risks to developed and emerging economies. Consider the following statistics across key markets like US & UK which highlight the urgency to infuse advanced analytics to prevent fraud and scams:

According to the FBI’s 2023 Elder Fraud Report, 101,068 complaints (age 60+) were reported in 2023, with a total loss of $3.4 billion. This is an increase of 15% from the prior year in terms of volume and 11% in total losses.
UK finance, a leading financial services trade association in the EU, reports in their 2023 annual report that fraud continues to be “one of the most concerning issues for society” and in the first half of 2023, “£580 million was stolen from victims of fraud by criminals”.

Challenges with Current Fraud Detection Systems

Fraud management executives and compliance officers face significant challenges in their efforts to improve fraud detection systems:

Real-Time Detection: Current fraud detection technology often fails to predict and prevent fraud in real-time.
Adaptability: Traditional rule-based systems cannot adapt to new or emerging threats.
Customer Impact: Many platforms are not able to detect scams such as authorized push payment fraud.
Data Integration: Existing systems often struggle to integrate and analyze data from multiple sources, limiting their effectiveness.

Financial Crime Technology Innovation

Wipro, a leader in the Financial Crime Compliance domain and a Premier AWS Partner for the past 10 years, has been helping banking customers navigate these challenges and has developed a deep level of understanding. With that expertise, Wipro collaborated with AWS to build an Intelligent Financial Fraud Detection (IFFD) solution, with the aim of filling in the existing technology gaps and platform limitations.

IFFD leverages best-in-class AI technology and advanced analytics, including deep learning and behavioral-based contextual analysis, to predict emerging threats and scams.
IFFD’s goal is to reduce fraud losses and prevent scams by forewarning the customers before funds are transferred out of the account.
IFFD exceeds industry expectations for identifying and preventing fraud in real-time, helping to solidify customer trust with the financial institution.
IFFD’s deep learning technology enables the system to learn from the historical transactions and predict emerging fraud patterns.
Unlike many third-party fraud platforms, IFFD provides model explainability, detailing how the models arrive at fraud decisions – a critical capability for auditability and regulatory compliance.
IFFD targets a false positives rate of less than 5%, by more accurately identifying true positives in real-time.
IFFD enables agile deployment of new fraud models as threats evolve, such as investment scams, business email compromise, romance scams, and grandparent scams.
IFFD can be deployed as a challenger model which can operate alongside and benchmark against the existing fraud detection systems.
IFFD can augment or integrate with the existing fraud management platforms, enabling a gradual modernization and transformation, rather than a disruptive rip-and-replace approach.

Case Study

Wipro is a premium partner for one of the top financial institutions in the US, involved in a multi-year program to modernize their financial crime compliance technology landscape.

The Challenge: Identifying Elder Financial Exploitation

The bank had identified a gap in their existing platform – it was unable to effectively identify and prevent instances of elder financial exploitation. This resulted in significant embarrassment and emotional distress for their elderly customers who fell victim to these scams.

Wipro’s Intelligent Financial Fraud Detection (IFFD) Solution

Leveraging our position as incumbent technology partner, we engaged with the bank to address this challenge through a proof-of-concept (POC) deployment of our IFFD solution.

The key differentiator of IFFD is its ability to analyze customer behavior and transaction patterns to predict the underlying logic and reasoning behind both non-monetary and monetary transactions. By applying deep learning models, IFFD can accurately risk-score each transaction variable, enabling the early detection and prevention of fraudulent activities – whether they be account takeovers (ATO) or impersonation scams targeting elderly customers.

IFFD sends real-time notifications to customers, allowing the bank to obtain direct confirmation from the account holder before processing potentially suspicious transactions. This proactive approach helps prevent both financial losses and the emotional distress caused by elder exploitation.

IFFD stands out with its unparalleled model explainability, providing the bank’s compliance teams and regulators with clear insights into the reasoning behind the platform’s decisions. This transparency is crucial for building trust and ensuring the responsible use of AI-powered fraud detection.

The Outcome: Protecting Vulnerable Customers and Enhancing Compliance

By using Wipro’s IFFD solution, the bank was able to overcome the limitations of its legacy fraud platform and establish a stronger defense against elder financial exploitation. By accurately identifying suspicious activities and empowering customers with real-time transaction verification, the bank significantly enhanced its ability to protect its most vulnerable clients.

The IFFD solution’s advanced analytics and explainable AI capabilities strengthened the bank’s overall financial crime compliance posture, providing regulators with the visibility and assurance they require.

Solution Overview

Wipro’s Financial Crime Compliance practice has extensive field experience. More than 3,000 team members have engaged on over 40 projects focused on anti-money laundering (AML), fraud, screening, know your customer (KYC), customer due diligence (CDD), and regulatory reporting for tier 1 and 2 firms. Wipro’s Intelligent Financial Fraud Detection (IFFD) solution was built on AWS to address common challenges through use of the following:

Advanced AI/ML Models: IFFD leverages AWS services like Amazon SageMaker to power machine learning and deep learning models that can more accurately identify fraud in real-time across a wide range of use cases.
Explainable AI: The solution provides transparency into model decisioning using techniques like SHAPley values. This helps foster trust and auditability.
Scalable Architecture: IFFD is built as containerized applications using Amazon Elastic Container Service (ECS), enabling the solution to rapidly scale compute capacity as needed to handle high-volume data streams.
Modernization Path: The IFFD solution is designed to augment and extend financial institutions’ existing fraud management platforms through secure APIs, helping modernize legacy systems.

Solution Architecture

The diagram below explains the end-to-end flow for the IFFD Solution for a user transaction.

Figure 1 – Solution Architecture

IFFD runs in a dedicated AWS account and makes a secure connection to the transaction data repository. For real-time decisioning, the banking application will need to call the IFFD endpoint.
The transaction details are sent to an API endpoint, backed by a highly available architecture
The transaction is handled by the orchestration service which runs on Amazon Elastic Container Service (ECS), which decides which steps to execute next.
The data can be hydrated with additional context from the case management database, which is stored in an Amazon Relational Database Service (RDS).
The orchestrator packages the transaction data and relevant context and calls the fraud detector service.
The fraud detector service runs machine learning inference via the custom model hosted on Amazon SageMaker. The inference result along with explainability data is returned to the fraud detector service.
The result is stored in the case management database.
The response is sent back to the requestor via the orchestrator service
The new case details are automatically extracted from the database by AWS Lambda and a request is made to update it in the integrated case management system.

Conclusion

Wipro’s Intelligent Financial Fraud Detection (IFFD) solution, built on AWS, represents a transformative approach to combat the escalating challenges of financial fraud and scams. IFFD leverages advanced AI and machine learning capabilities to predict and prevent emerging threats in real-time.

Through its behavioral analysis, real-time transaction monitoring, and explainable AI features, IFFD has demonstrated the ability to significantly enhance fraud management capabilities. IFFD has demonstrated the ability to detect and prevent elder financial exploitation scams that other platforms had failed to identify. It has also been able to reduce false positive rates by over 95%, minimizing customer friction.

By seamlessly integrating with financial institutions’ existing fraud management infrastructure, IFFD enables a gradual modernization pathway, rather than a disruptive rip-and-replace approach. Wipro’s deep domain expertise in financial crime compliance, combined with the scalability and reliability of the AWS cloud, makes IFFD a compelling solution for banks and other financial services organizations seeking to stay ahead of the evolving fraud landscape.

.

.

Wipro – AWS Partner Spotlight

Wipro is an AWS Premier Tier Services Partner and Financial Services Competency Partner that harnesses the power of cognitive computing, hyper-automation, robotics, cloud, analytics, and emerging technologies to help clients adapt to the digital world.

Contact Wipro | Partner Overview | AWS Marketplace

AWS Partner Network (APN) Blog