Artificial Intelligence & Machine Learning , Finance & Banking , Fraud Management & Cybercrime

Fake Paystubs Are Draining Billions From Lenders

How AI Could Solve Failings of Traditional Employment and Income Verification
Fake Paystubs Are Draining Billions From Lenders
Image: Shutterstock

Thousands of loan applications have listed a small Baltimore computer store as the place of employment for borrowers claiming high salaries. In reality, the firm never employed that many people. The company allowed individuals to claim it as an employer in exchange for cash. As a result, lenders received thousands of loans with bogus employment information, and their outdated verification systems did not detect them.

See Also: Maximizing data utility in mission delivery, citizen services, and education

Income and employment questions are basic tenets of applications for credit, insurance, rentals and other agreements. But a growing number of fake paystub providers are generating phony documentation and answering phone calls to verify employment, even if the whole company is fake. Lenders face high costs and limited coverage of HR databases for employment verification, and fraudsters are exploiting these gaps.

According to Point Predictive's 2024 Auto Lending Fraud Trends Report, income and employment fraud is a pervasive problem, accounting for nearly half of the industry’s $7.9 billion fraud loss exposure. In 2023, income and employment fraud accounted for a massive $3.6 billion in losses for auto lenders, representing a hefty 45% of the total fraud-related losses in the industry.

Employment verification fraud affects a wide range of lenders. These loans are more likely to default, increasing charge-offs and financial strain on lenders.

What Isn't Working

The biggest pain point in the verification process is the lack of real-time, holistic data access that lenders need to make accurate, confident decisions. Ronan Burke, CEO and co-founder of Inscribe AI, said that even with automation, many financial institutions are still dealing with "fragmented data sources, creating gaps in the information they rely on."

The problem is not new, but with advancement in digital banking, submission of paystubs has become a cumbersome process. Verifying income and employment has become a complex and costly challenge for lenders due to several key issues.

Document forgery is a major concern, as 1 in 10 paystubs submitted contains hard-to-detect forgeries. Analysts often face review fatigue from the sheer volume of paystubs, leading to missed information. Also, employment databases are incomplete, covering only about 45% of applicants, which forces lenders to rely on alternative methods such as bank statements or manual verifications, according to Point Predictive's Cracking The Code - The Evolution of Income and Employment Validation.

These verification methods not only increase manual workloads but also lead to loan application abandonment because of the additional documentation requirements placed on consumers. Providing a genuine paystub is a slow and cumbersome process for a consumer. But providing a fake paystub is much easier. It can take as little as 90 seconds to produce a fake paystub using one of the more than 12,000 paystub generation sites available on the internet.

While paystubs are notoriously unreliable, the most commonly accepted alternative, employment verification, is costly and it can be time-consuming. Centralized human resource databases offer the convenience of instant employment verification, but typically at a high cost. Each verification 'hit' can cost $30 or more. Verification has become even more challenging with the rise of the gig economy, which introduces nonstandard income and employment streams. Human resource databases often only have information on 30% to 45% of applicants, leaving a majority of applications uncovered, according to Point Predictive.

Emerging credential-based solutions can connect directly to payroll providers or bank accounts, but adoption remains low, as only 3% to 5% of consumers use them.

“All of these challenges create friction in the verification process, leading many potential borrowers to abandon their applications due to the burdensome and time-consuming requirements,” said Frank McKenna, co-founder at Point Predictive.

AI: A Game Changer?

Income and employment verification requires a multifaceted approach. Lenders must create frictionless solutions that eliminate the need for consumers to provide extensive documentation. But they also can use data validation and automation with artificial intelligence to replace manual paystub reviews, which are often prone to human error and inefficiency. AI-driven solutions can assess the accuracy of income claims more effectively.

Finally, expanding the coverage of verification systems to address 80% to 90% of consumers would enable lenders to process applications faster and with fewer manual interventions, leading to a more efficient and reliable system overall.

Many solution providers including Point Predictive, Plaid, Atomic, TruWork, Inscribe AI and Ocrolus are tackling these challenges in different ways. For example, Point Predictive's IEVaidate uses its proprietary data covering more than 260 million income reports and seven other sources of data to identify, in the background, if borrowers are being truthful about their income, using data validation instead of manually reviewing paystubs. Plaid verifies bank statements and links with payroll connectivity, while Inscribe AI uses artificial intelligence to detect fraudulent income documentation.

Inscribe AI uses advanced AI and machine learning models to streamline the income and employment verification process. Automating tasks such as document analysis and fraud detection allows for identifying patterns and anomalies that human reviewers might miss, reducing the risk of accepting falsified documents. Inscribe AI's solution cross-references data from multiple sources, including bank statements and paystubs, to build a comprehensive profile of the borrower, helping lenders make more informed decisions in real time.

This depth of analysis, combined with real-time insights, sets Inscribe apart from traditional systems that often miss subtle but critical fraud signals.

Industry leaders are also betting big on automation. By automating much of the verification process, banks can drastically reduce the need for manual labor, which has historically been both time-consuming and expensive. With AI tools handling tasks such as document analysis, fraud detection and cross-referencing data in real time, lenders will not only be able to cut down on operational costs but also avoid losses for bad loans, according to Ronan Burke, CEO and co-founder Inscribe AI.

One of the biggest areas of savings is in the reduction of manual work. Faster processing times mean that banks can handle more applications in less time, which can directly improve the bottom line. They also accelerate loan approvals, enabling banks and lenders to generate revenue more quickly.

Glenn Kranis, director of Auriemma Roundtables, said multiple banking fraud practitioners are exploring OCR tools that can help detect misrepresentation in proof of income documents by identifying discrepancies and alterations in scanned records. “Specialized tools for self-employed applicants could aid in decision-making by evaluating their income through alternative documentation,” Kranis said.

The use of one-time passcodes in the verification process could enhance security by linking lenders to official communications from verified employers or financial institutions to ensure the legitimacy of employment or income claims, he said.

"You want it to be frictionless, but fast, economic and accurate," said Steve Lenderman, head of fraud solutions in North America at Quantexa, adding that it is similar to onboarding a new employee. "The job market is too tough and competitive to tell a candidate, 'I will get back to you in four weeks.' The backlash is: If you can't verify a legit employee, you may lose out on a good one."


About the Author

Suparna Goswami

Suparna Goswami

Associate Editor, ISMG

Goswami has more than 10 years of experience in the field of journalism. She has covered a variety of beats including global macro economy, fintech, startups and other business trends. Before joining ISMG, she contributed for Forbes Asia, where she wrote about the Indian startup ecosystem. She has also worked with UK-based International Finance Magazine and leading Indian newspapers, such as DNA and Times of India.




Around the Network

Our website uses cookies. Cookies enable us to provide the best experience possible and help us understand how visitors use our website. By browsing databreachtoday.com, you agree to our use of cookies.