Time to Automate All Bank Processes

The uncertain economic environment, with a recession likely on the horizon and inflation driving up costs, has given banks a unique opportunity: revisiting their existing compliance and operational systems, and exploring long-term, scalable solutions in response to looming and increasing regulatory pressure.

Leveraging machine learning and automation to power digital transformation can address the concerns that keep bank directors up at night — especially since financial institutions may be expected to begin providing more data over the coming months. This comes at a time when banks are dealing with a number of external challenges; however, bank directors know they cannot skimp on adherence to strict compliance requirements. Missing a revenue goal is unfortunate, but from what we’ve heard from our customers, missing a compliance requirement can be a devastating blow to the business.

Increasing Regulatory Risk
Banks and other lenders may encounter financial strain in adding more compliance staff to their teams to address new regulations. Among them, Section 1071 of the Dodd-Frank Act requires financial institutions to report demographic information on small business loans. Regulators are reworking the Community Reinvestment Act. In response, banks are considering how they can leverage automated compliance systems for fair lending, loan servicing and collections.

Bankers are quick to acknowledge that the manual processes involved in data verification should be eliminated if their institutions have any hope of staying ahead of the curve. Furthermore, labor shortages and increased competition for talent has increased costs associated with these tasks — yet their necessity is imperative, given regulatory scrutiny.

As loan originations decrease during an economic slowdown or recession, it may look like delinquency rates are increasing as the ratio of delinquent loans to originations increases — even with no notable changes in delinquency cases. The increasing ratio could trigger scrutiny from the regulators, such as the Consumer Financial Protection Bureau.

If that happens, regulators look into whether the borrower should have received the loan in the first place, along with any fair lending bias concerns, and whether the bank followed appropriate procedures. Regulators will scrutinize the bank’s loan servicing and collections compliance procedures. Given that traditional manual reviews can be more inconsistent and vulnerable to human error, this becomes an incredibly risky regulation environment, especially where data integrity is concerned.

To mitigate risk and increase operational efficiency, banks can use end-to-end document processors to collect, verify and report data in a way that adheres to existing and pending regulation. Implementing these processes can eliminate a large portion of time and labor costs, saving banks from needing to recruit and hire additional compliance professionals every time fair lending and servicing requirements become more demanding.

Automated Processing
Lenders like Oportun, a digital banking platform powered by artificial intelligence, have found that leveraging intelligent document processing has reduced the cost of handling physical documents and traditional mail by 80%, increased margins, lowered instances of human error and improved data integrity. Enhancing customer experiences and providing quality data are crucial for Oportun; this makes their operational goals more cost-effective and scalable, and increases the capacity for Oportun’s team.

“[Automation] has helped us establish some strong controls around processing mail and servicing our customers,” Veronica Semler, vice president of operations at Oportun, says. “It’s reduced the risk of mail getting lost … it has increased our efficiency and made things easier for our team members in our stores.”

Institutions that leverage automated systems and machine learning for compliance can reduce labor costs, provide customers with high quality, efficient service and deliver accurate data to regulators. This provides companies like Oportun, which was an early adopter of machine learning, with an advantage over competitors that use traditional manual review methods.

Implementing document automation into existing systems allows banks to address compliance concerns while laying the groundwork for growth. Automation systems provide the tools for banks to reduce friction in lending and operations, enhance their controls and reduce human error — giving boards confidence that the bank can provide accurate, quality data ahead of any new fair lending and servicing regulations. Now is the time for boards and executives to recession-proof their banks and facilitate long-term success by investing in automation for document processing.

5 Best Practices for Digital Identity Verification

Attacks on the financial sector have increased steadily for two decades, and the volume of reported attempts surged in just the last few years.

In fact, 68% of financial services providers reported an increase in fraud attempts compared to the prior year. Fraud in the account opening processes is endemic; in response, institutions are using multi-layered verification to locate, approve and onboard legitimate customers with low friction while deterring fraud and maintaining compliance. A robust identity verification program allows platforms to capitalize on digital adoption while delivering a seamless customer experience. Fifty-three percent of Americans report that being prompted to take extra steps to verify their identity makes them trust that company more. And those who report being less trusting are less likely to engage in desirable downstream business practices.

A lack of trust creates a drag on profits while compromising the end-user experience. But institutions can use several best practices to locate and approve new legitimate customers, significantly lessening friction or fraud and streamlining the customer journey.

1. Analyze Multiple Layers of Data
Forty-five percent of organizations say they perceive multiple layers of identity attributes as a best practice. As fraudsters increasingly add sophistication to their schemes, additional layers, or “blankets,” of attributes that work together are the key to a seamless customer experience and fraud mitigation. Solutions that orchestrate multiple dynamic data sets not only detect and deter fraud — especially synthetic identity fraud — but don’t add friction because the solution is predicated on data collection practices that are easy to explain and defend.

Multiple layers at the heart of the identity verification process identifies legitimate customers more quickly and accurately, and uses additional verification methods only when absolutely necessary.

2. Layer Machine Learning with Human Fraud Expertise
Financial service providers can balance user experience with identity verification standards by combining  increasingly adopted technologies with human fraud expertise. Financial institutions have the power to analyze massive amounts of digital transaction data by applying supervised machine learning (ML) to the identity verification process, creating efficiencies by recognizing patterns that can improve decision-making.

Coupling this with human expertise and intuition gives institutions the best of both worlds: enhanced anti-fraud protocols and new, more usable data sets that improve identity verification efforts going forward. Machines are great at detecting trends that have already been identified as suspicious, but have a blind spot of detecting novel forms of fraud. It’s critical that providers layer human fraud expertise on top of machine learning.

3. Embrace Data and Decision Transparency
Many ML-based solutions provide a pass or fail score that is as opaque as it is simple. Without visibility into decisioning data, institutions are left to depend on restrictive and hazy score-based identity proofing models. These “black box” solutions don’t offer data intelligence visibility; instead, they apply common engine logic across multiple customers and industries.

An effective identity verification solution should provide a continuous data feedback loop so institutions can understand and explain to regulators and consumers why they made certain decisions. This allows financial institutions to better assess their risk and fine-tune the identity verification processes to best fit their needs. This is nearly impossible to do with a system that relies on “black box” algorithms and little governance of modifications from one company to another.

4. Implement Customized Identity Verification Workflows
The ability to customize identity verification settings to meet specific customer needs is quickly becoming mission-critical. Every organization is different; every financial institution has different verification protocols that reflect these unique needs. This includes the ability to tweak and tune identity verification settings in real time, without the help of IT. Every institution needs the ability to act quickly as they anticipate attacks, adapt to changes in human behavior and respond to the emergence of new customer segments, profiles and needs.

At the same time, institutions need to empower decision-makers to collect less sensitive information or enact pre-qualification formats for certain applications, streamlining customer onboarding without compromising identity verification standards.

5. Cross-Industry Fraud Intelligence
It’s common for fraudsters to jump from industry to industry as they carry out their plans, which means that effectively fighting fraud is a group effort. With the right identity verification solution in place, financial institutions will have visibility into serial, multi-industry fraud schemes and trends and data across industries and channels.

As the financial sector moves towards a post-pandemic reality, fraud attempts are likely to grow alongside customer expectations. Identity verification will be an operational necessity and a moral imperative, keeping financial institutions and consumers safe in a challenging digital environment.

The Future of Fighting Financial Crime

The anti-financial crime landscape is continuously evolving, and financial institutions need to stay a step ahead of emerging fraud trends and regulatory compliance challenges to protect their customers and themselves from loss and reputational damage.

As consumers become increasingly reliant on the speed and convenience of digital banking products, institutions should consider end-to-end financial crime management solutions that offer real-time fraud detection, targeted AML transaction monitoring and automated regulatory reporting to fight financial crime and strengthen compliance. With artificial intelligence, including machine learning and robotic process automation (RPA), behavior-based anti-financial crime management solutions can help institutions increase the effectiveness and efficiency of their fraud and AML programs.

But artificial intelligence relies on the power of big data. Anti-financial crime management solutions need an immense data set from multiple sources, including core, ancillary, open-source, third-party and consortium data. Artificial intelligence can be applied to this data with cross-institutional analysis in a cloud-based environment. Solutions built with big data and artificial intelligence reduce false positives and increase the quality and accuracy of alerts.

Analytical agents built with machine learning algorithms continuously analyze data to improve analytical performance. A large, cross-institutional data set in a cloud-based environment allows machine learning agents to train on labeled data from thousands of institutions, achieving performance levels that cannot be matched by a single institution with a limited, isolated and restricted data set. Machine learning can significantly improve analytical performance, helping institutions reduce false positives and reduce the alert review time to increase the efficiency of investigations, while continuing to detect new and emerging criminal trends.

Machine learning agents use mathematical and statistical models to learn from data without being explicitly programmed. These agents analyze new data, including transactions, demographics and customer behavior, and utilize this evidence in transaction monitoring alerts. The alerts provide feedback to the cloud-based data set, where the training data, which includes the transaction, demographic and customer-behavior evidence, and labeled data, such as cases, marked transactions and return items, are input to a machine learning algorithm.

This data is used to train many different types of machine learning agents to determine which type of agent performs best for a particular typology. Before training an agent, the data is split into training, testing and validation data sets, so that the results of the training can be validated in an unbiased manner. Anti-financial crime management solutions can use precision, recall and false positive rate to validate analytical performance and assign the most suitable analytical agent to a particular fraud or money laundering typology.

Strengthening Processes with Robotic Process Automation

Institutions can leverage technology such as RPA to improve internal processes, strengthen anti-financial crime management programs and ensure regulatory compliance. Using RPA to improve workflow automation can save financial crime investigators time by reducing manual tasks, automating steps in alert triage and regulatory reporting processes, and through prepopulating and submitting reports with consistency, speed and accuracy.

Enhanced data collection through RPA reduces human error that occurs during manual data collection and transference. It also automatically and reliably integrates multiple data sources in your anti-financial crime management and compliance solutions.

Anti-financial crime management programs can use RPA agents to validate information populated in a currency transaction report or a suspicious activity report, automatically submit reports, intelligently package related alerts together, and automatically assign work to a team or an investigator. Such solutions can also automatically triage alerts and segment customers into appropriate risk categories, increasing the efficiency and effectiveness of financial crime investigations.

To stay ahead of financial crime trends, financial institutions should consider the benefits of cloud-based solutions that leverage artificial intelligence to increase the effectiveness and efficiency of anti-financial crime management and compliance programs.

A Deep Dive Into Wire Fraud and Business Email Compromise

Consumers demand for fast and convenient payments channels has increased opportunities for fraudsters to target financial institutions and their customers.

With wire fraud and business email compromise (BEC) attacks increasing, it is critical that banks remain vigilant to prevent fraud losses and reputational risks. We are sharing unparalleled data-driven insights into the current fraud landscape that we uncovered through the Verafin Cloud, with a deep dive into wire fraud and BEC. The Verafin Cloud contains an immense set of anonymized data from over 3,000 financial institutions, comprising $4 trillion in assets. Importing core, ancillary, open-source, third-party and consortium data, and analyzing over a billion transactions a week in the Verafin Cloud, we can accurately identify emerging fraud trends and create a substantial set of labeled fraud data to train machine learning analytics for fraud detection.

The Main Target for Wire Fraud
Criminals are constantly searching for weaknesses in banks’ wire fraud controls and will shift tactics to target points of least resistance – often your own customers. Criminals have refocused their efforts to leverage your customers as an attack vector, targeting them with known fraud scams. Statistics from the Verafin Cloud show that nearly three-quarters (74%) of all wire fraud cases targeted individuals, with elderly persons accounting for 63% of all people victimized by wire fraud.

BEC Behind Majority of Loss
While individuals were more frequently targeted by wire fraudsters, data in the Verafin Cloud shows that businesses sustained 73% of all financial losses to date, driven largely by BEC schemes. While most BEC attempts in our analysis involved wire transactions, 24% of BEC occurrences involved ACH transfers, demonstrating this channel is not immune to attack. A high value, high speed, and widespread scheme, BEC has become the No. 1 reported crime to the FBI, and is an ever-increasing threat to all banks.

Payee Risk Analysis
At many banks, a wire sent to a first-time beneficiary is automatically considered high risk. This assumption creates undue friction for your customers, as well as massive alert volumes — especially when a large proportion of wires from banks are destined for new recipients. This figure was substantial in our data: 23% of wire transfers were directed towards new payees for a customer. Banks should consider technology that provides visibility into the transaction counterparty in real time to ascertain whether a wire recipient is truly suspicious or has a trusted history of activity at other institutions.

A Step Ahead
Wire fraud is a growing threat for financial institutions. As fraud schemes evolve and become more sophisticated, wire transfers —which can be high value and irrevocable — are the perfect target for fraudsters. As criminals increasingly target your customers with a variety of fraud scams and schemes, banks must remain vigilant and ensure that holistic fraud detection and management solutions are in place to prevent loss and stay a step ahead of financial crime.