Artificial Intelligence, Data

‘This year, banks will strive to balance two opposing forces’: The state of AI in banking 2022

  • AI developments in banking have so far been restricted mostly to back-end uses. This year, there is a desire among service providers to focus on innovating more for the front-end.
  • As AI becomes smarter and banks begin holding increasingly intimate data about their customers, the industry is expected to progress slowly and responsibly.

Email a Friend

‘This year, banks will strive to balance two opposing forces’: The state of AI in banking 2022

McKinsey estimates that AI and analytics have the potential to generate an incremental value of up to $1 trillion for global banking each year. 

The use of AI in banking is not new and can be traced back to the 1980s. That’s when specialized computer programs called expert systems rose to prominence. The technology was developed to mimic the decision-making of a human expert. Expert systems were initially deployed in finance to predict market trends and provide customized financial plans for customers. 

With that in mind, what do the AI developments in banking look like as we head deeper into 2022?

Focus on personalized front-end experiences

Most AI deployments run on the back end of financial services. Common use cases include predictive analytics, cybersecurity, fraud detection, and credit scoring. On the front end, use cases for AI that interact directly with the customer have been limited to customer service chatbots and robo-advisors. There is a desire in the banking industry to bring AI closer to the end-user by creating personalized app experiences. Analysts use the term ‘hyper-personalization’ to describe the next stage in digital banking.

Consumer authentication emerged among the first use cases of AI in banking visible to users. Today, AI has helped introduce behavioral biometrics to authenticate users, which include features like speaker recognition (authenticating individuals by the way they talk), motion/gait (authenticating individuals by the way they walk or move), and keystroke dynamics (authenticating individuals by the way they type).

“Customers may not see AI technology at all, but those that do will see it in the context of a better customer experience,” said Scott Zoldi, chief analytics officer at FICO. “Banks are starting to use AI to better understand the totality of individual customers' financial relationships, lifestyles, and histories.”

On the front end of digital banking, AI’s prime usage will see consumers receive more intelligent, personalized, and valuable services. To this end, fintechs may use AI to analyze customer data in real-time and use the analyses to improve customer experience. 

An example is the automatic couponing platform Honey, which was acquired by PayPal in 2019. Honey’s technology uses markers like cookies, pixel tags, and web beacons to show customers tailored deals based on their preferences. Its automatic couponing feature puts the buyer first, by waiting for them to choose what they want to purchase. Once chosen, Honey runs all eligible promo codes at checkout and automatically applies one that saves the most money on the buyer’s behalf.

“We’re going to see AI and machine learning play a larger role in fintech, and help embedded finance tools become smarter and provide more value to consumers,” Randy Kern, CTO at Marqeta, told Tearsheet.

Kern believes that payment service providers find themselves in a leveraged position to use AI to create greater value for customers. Sitting atop the payment journey, card issuers see a larger part of the payments ecosystem than merchant acquirers. This allows them to use data to improve the consumer experience. For instance, issuers can help their business customers make proactive recommendations to help cardholders save money.


“By leveraging AI, companies can be more agile and offer different incentives based on consumer behavior,” Kern said. “Companies that nail the customer experience with payments, while protecting them from bad actors, will stand apart from the competition.”

Continued development on the back end

As bank branches continue to close, customers are increasingly banking digitally. Behind the scenes, AI is being used to service that shift in consumer behavior. 

In 2022, fraud remains rampant in banking and will continue to be a focus of AI development.

“The most popular use cases for AI in customer banking will continue to be fraud detection, underwriting, and risk management,” John Whaley, svp of emerging products at Prove, told Tearsheet. “I think we'll see more sophisticated banks adopt AI enabling them to evaluate whether a user is legitimate or a fraudster based on individual behaviors.”

The Federal Trade Commission disclosed that consumers reported losing more than $3.3 billion to fraud in 2020, up from $1.8 billion in 2019. AI’s ability to comb through a large amount of data quickly in real-time and to detect anomalies in patterns could potentially catch fraud ahead of time.

An example of such use is Fiserv’s AI-based fraud detection solution for card issuers, called Advance Defense. It is designed to help issuers minimize fraud losses. The tool uses data from the client card issuer and a consortium to identify fraud patterns at the merchant ID and instantly recommends tailored fraud rules. Banks otherwise spend days or weeks analyzing fraud data to develop prevention strategies.

Machine learning models have been used to track and understand patterns in transactions, which can tell a fraudulent transaction apart from an authentic one. As fraudsters get creative, these models have also displayed their ability to catch and learn new methods employed to commit fraud. Upon the detection of fraud, an AI system can be utilized to either automatically reject the transaction or flag it for investigation.

“Today, businesses typically apply detection techniques to catch fraud after it occurs, a process that can be time-consuming and compute-intensive due to the limitations of today's technology, particularly when fraud analysis and detection is conducted far away from mission-critical transactions and data,” said Anthony Lipp, partner and global financial services strategy leader at IBM. 

“Due to latency requirements, complex fraud detection often cannot be completed in real-time – meaning a bad actor could have already successfully purchased goods with a stolen credit card before the retailer is aware that fraud has taken place.”

The ethical side

Any use of AI, whether to create new personalized products or for compliance, inevitably requires banks to use customer data. Such concerns are not lost on banks. Service providers understand the need to balance out the tension between using AI to boost customer engagement and to avoid overstepping the ethical aspects of the technology, like biased credit models and influence over user behavior.

“2022 is a year in which banks will strive to balance two opposing forces. One is the focus on more customers engaging with their bank digitally, sharing more data, and expecting a more customer-centric decision to be made with AI. The opposing force is the banks' increased awareness of Ethical AI and the fact that many types of AI/ML can be unethical and even dangerous to utilize,” said FICO’s Zoldi.

As banks collect and record intimate and personal data about their customers, their responsibility to act as per privacy laws is becoming central. Consumers often have little idea about how and what data of theirs is being used. Explanations and permissions are often buried in the terms and conditions that many people do not read.

In a recent survey, consumers, when made aware of how fintechs use their data,  desired more control over how their data is being used and where.

Another ethical concern raised about AI is in customer screening. The technology may serve to reinforce unconscious biases and prejudices. 

Zoldi believes that a “responsible AI model development governance framework” is at the heart of the efforts to deploy the technology. “The industry will likely focus on slow progress toward Responsible AI instead of rushing in with AI that ultimately may have to be backed out or could harm people with biased outcomes,” he said.

The winners and the losers

The winners in financial services will be those that digitize the fastest, using technology to get un-stuck on manual processes done by humans, according to IBM's Lipp.

Today, the banking environment is becoming more ecosystem-based, with a variety of product and distribution partners. It is hence becoming critical that platforms have the ability to integrate with other platforms. “The winners will be those that have gone through extreme digitization and transformed their data environment to seamlessly integrate across different platforms,” said Lipp. “The losers will be those that haven't made that transition yet – focusing on the AI tools but not thinking through how to effectively use them.”

Experts believe that end users will also win with increased AI in banking, especially those seeking personalized experiences. 

“Individualized customer decisioning will be a big benefit for consumers who want a unique and personalized mix of services,” Zoldi said. 

Analytical data about consumers’ financial health can enable them to understand where they financially stand more accurately. Additionally, services like AI assistants, AI advisors, and scenario modeling can be used by customers to see the impact of potential actions on meeting their financial goals.

0 comments on “‘This year, banks will strive to balance two opposing forces’: The state of AI in banking 2022”

Data, Lending

The Financial Data Evolution: Unraveling lending in open finance with Akoya’s Rishi Kapadia

  • As part of Tearsheet Talks: Lending x Credit, Data, Rishi Kapadia, product manager at Akoya, discussed "The Financial Data Evolution: Unraveling Lending in Open Finance."
  • Kapadia sees open finance facilitating an evolution in the whole user experience in the lending process.
Tearsheet Editors | October 31, 2023
Data, Lending

Best practices for fusing internal and external data to enhance credit decisioning for SMB lending with Stripe’s Yaakov Erlichman

  • Learn strategic and tactical best practices for maximizing the fusion of internal and external data to enhance credit decisioning and risk management for SMB lending.
  • Stripe's Yaakov Ehrlichman, head of capital and SMB risk, covers both conceptual and implementation considerations for reconciling data sources tailored to the unique needs of the SMB market.
Tearsheet Editors | October 26, 2023
Data, Partner

How payments providers can tap into the growing SMB market 

  • The POS market continues to grow, driven by digitalization and new vertical offerings, but there's a lack of available data around SMB revenue, growth trends, and other metrics to help capture these businesses as customers.
  • Data firm Enigma provides comprehensive data about the identity and financial health of millions of businesses, enabling visibility into real revenues and financial trends across this growing market.
Enigma | August 24, 2023
Data, Keeping the bad guys out

The network approach to fighting fraud

  • With fraud evolving, prevention and detection mechanisms are changing as well.
  • FIs and technology providers are turning towards approaches where critical data about fraudulent activities can be shared with FIs and businesses in the ecosystem via a network approach.
Rabab Ahsan | July 05, 2023

As tides shift against screen scraping, complications abound for smaller FIs

  • Screen scraping is experiencing increased scrutiny from regulators and competition from API-based solutions. The practice may finally be on its way out.
  • But what would an API-based data sharing environment look like? How will the transition away screen scraping occur and would it be fair to all institutions?
Rabab Ahsan | June 09, 2023
More Articles