Artificial intelligence

As use of Bank of America’s virtual assistant Erica balloons, the bank is continually fine-tuning and feeding the tool new information to better respond to customer inquiries.

That’s according to Nikki Katz, Bank of America’s head of digital, who oversees the team that supports the bank’s mobile app and website, as well as AI-powered Erica. 

Since Erica’s 2018 launch, the bank’s virtual assistant has been used more than 2 billion times, and it’s responded to 800 million inquiries from about 42 million clients, the bank said in a release in April. In light of its popularity with clients, Bank of America has sought to integrate the tool more deeply into the bank’s digital functions. 

Katz declined to comment on how much the Charlotte, North Carolina-based lender has invested in Erica to date. The bank plans to spend $3.8 billion on technology initiatives in 2024, CEO Brian Moynihan said last October. Bank of America aims to take a thoughtful, controlled approach to AI while searching for more areas where Erica can provide benefit, Katz said Thursday during an interview.

“Services like Erica — and, in general, all of our digital experience — are a big part of why people stay once they're here,” Katz said of client satisfaction. “It is definitely a big part of retention.”

Editor’s note: This interview has been edited for clarity and brevity.

BANKING DIVE: Why do you think Erica is resonating with the bank’s customer base?

NIKKI KATZ: Erica is really almost a gateway to client servicing, connecting clients to products and services, making it easier to access those products and services, simplifying and basically connecting life and financials for them. Erica can provide a weekly snapshot of your month-to-date spending, facilitate scheduling in-person meetings, flag a recurring charge and help clients understand what they need to pay attention to, to adjust their spending patterns.

There’s two ways that clients engage: inquiring of Erica — so, asking Erica questions — and getting insights from Erica, so Erica’s going to them.

We always anticipated, when we first launched, that people would ask Erica questions, but more and more, that has shifted, and now 50% of Erica interactions are people taking advantage of those insights. That’s things like monitoring recurring subscription fees, looking at the spend path, managing and monitoring deposits, refunds. Clients are asking for their account or routing number, instead of having to root around for it. Erica searches across your account, making it easy to look for a particular transaction. People also use Erica to assist with money transfers, bill pay.

How does Bank of America blend digital and human response to client needs?

Clients are doing more and more digitally, and we want to be there for that. But in the key moments that matter, when something's truly sensitive, or where you really need that human interaction, having branches and our high-touch strategy is a differentiator. It's about striking the right balance between technology and human channels.

We did 23.4 billion interactions with our customers via digital channels last year. Well north of 90% of all of our client interactions are now flowing through digital channels. Customers are using those channels more frequently; it’s becoming a habit.

For well over 90% of interactions, digital is perfect. If a client truly prefers to talk to a human, we definitely don’t want to be in the way of that. In that moment, Erica can essentially expedite that interaction, ensuring you get into the right queue, with the right specialist.

What kind of risks do AI-powered tools like this present? 

Our approach to AI is kind of controlled AI, because it has to be, in the space we’re in. There’s no room for an AI to hallucinate when dealing with personal finances. We are very thoughtful about where we apply AI. All the responses Erica produces, everything goes through really rigorous review to ensure she’s not ad-libbing. She’s not telling you something we wouldn’t stand behind.

This is an area where we have a ton of investment, we have a huge patent portfolio, we have a huge pipeline of AI-based projects we’re working on. This is an area we believe has a ton of value and opportunity for our industry, but we’re definitely going to come out on the more thoughtful side because of risk.

What else is in the bank’s pipeline of AI-based projects?

There’s a lot of applications we’re looking at, both in terms of internal efficiencies, internal tools as well as client facing. Nothing that we’re able to talk about beyond Erica right now, but it’s definitely an area of investment and investigation for us.

What’s next for Erica?

We are constantly evolving Erica, every single day. We’ve made 50,000 different tuning adjustments to Erica since it was launched, and we continue to do that.

We’re constantly looking at our natural language model, and ensuring it’s understanding what’s being said to it. During the COVID-19 pandemic, when the Paycheck Protection Program came out, all of a sudden Erica was getting all these questions around PPP. This was a new program, and there was nothing already in the system for it, so we had to really quickly adjust and improve the knowledge base so Erica knew what to do with those types of questions.

Erica was born on the consumer retail side, but once we started bringing in wealth customers and retirement, we had to train Erica on how to talk about retirement products, how to talk about wealth and what questions come with that, what clients need around those domains.

We're also expanding, so now that same data and technology is powering our CashPro Chat on the business side. You're going to see that kind of expansion of Erica and that technology across more of our businesses, capabilities. We’re looking at opportunities internally as well as client facing, to leverage that. Erica now powers the search at the top of the mobile app. 

Small businesses and financial institutions, including fintechs, can benefit from leveraging artificial intelligence technology like machine learning in making lending decisions, a white paper report commissioned by Uplinq, a credit assessment platform for small-business lenders, found.

Together with using alternative data, adopting new technologies helps financial institutions offer advantages to their customers such as credit assessment and more efficient collections and loss mitigation, according to the white paper developed by Cornerstone Advisors, a management and technology consultancy for banks, credit unions, and fintechs.

“Data is noisy, and in other words, sometimes data is falsified. Sometimes, data is incomplete. Sometimes it’s not intentionally wrong; it’s just been reported incorrectly,” Patrick Reily, co-founder of Uplinq, told Banking Dive. “And so, the AI does a very good job of looking at the information integrity, and understanding when that information is reliable or when it may be suspect.”

But it is crucial to inspect and sort the data before using it for underwriting purposes, he noted. The research shows that financial institutions that used Uplinq-based underwriting saw higher small business loan approvals among protected categories — up from 38% to 82%, according to the press release.

Banks and lending institutions nationwide are continually trying to streamline underwriting and loan processing methods through all the available digital and AI solutions “to stay competitive and on the cutting edge of customer experience,” Stephanie Dunn, head of SBA lending at Grasshopper Bank, told Banking Dive via email.

“Innovation and soundness [are] every bank’s priority,” she noted, adding that the “emphasis is now also both on the lending institutions and the small business owners combined to be as informed as possible given AI data is now so widely accessible. Knowledge is power as we all know and AI has allowed more widespread access to knowledge than ever before in history.”

Though AI can play an important role in providing affordable loans to small businesses, the white paper cites studies that say women and small-business owners of color are often labeled “higher risk” and struggle to access credit from traditional lenders based on conventional underwriting criteria. This can push entrepreneurs to go for less advantageous funding sources like loans at higher rates from non-bank sources, according to the study.

Most traditional underwriting systems rely on human interactions, like people reading reports and making decisions. If they look at the same data repeatedly, humans sometimes fail to spot the discrepancies. Still, technology like AI stays put and helps differentiate between real and fake data, Reily pointed out. This helps to make lending decisions faster and reach more people, he said.

“You say, I’d like to ask for money, and then you have money in your bank account — that’s real empowerment. And so hopefully what it shows you is that fintechs can absolutely play in that space and they can exist in that space,” he noted. “Fintechs are always going to be challenged for cheap capital because they don’t get to take deposits. And so, they have to look at opportunities where the market will tolerate a higher interest rate or a higher cost structure. But as long as you’re focused on those segments, there’s room to play.”

Challenges in SMB lending

According to Dunn, the major gap that has long hindered accessing affordable loans is financial education. However, there is a proactive focus on providing loans to all businesses, including diverse markets across the country.

“Common barriers for SMBs across the country and more specifically in underserved communities include access to capital and how to efficiently bridge the business need with the bank’s actual loan products. The gap has always been and is still today, financial education,” Dunn noted.

But she pointed out that the lending landscape is very strong across various platforms, indicating that the “U.S. entrepreneurial spirit is active and thriving,” she said.

“Fintechs are challenged with finding streamlined processes in order to sustain their cost of funds. I would say both banks and fintechs are equally challenged with their respective efficiency metrics while maintaining credit quality and customer experience,” Dunn added.

Sponsored content

By ServiceNow

Every bank and credit union, regardless of size, needs to have a Generative AI strategy​ that enables them to succeed and grow. 

Remember RAP? There was a time when those regulatory accounting principles were new and fancy. AI is this generation’s transformation but with far more upside. While many industries are worried that AI will kill jobs, it’s not the case in financial services. In fact, Gen AI will help employees work smarter and more efficiently, freeing up staff members to do more critical and valuable work, such as focusing on acquiring and onboarding customers. It can also provide frictionless servicing, and help manage risk and compliance while modernizing technology, all things that lead to more productive and satisfying work and happier end users.  

Consider this: Gen AI could have a significant impact on the banking industry, boosting value from increased productivity of 2.8 percent to 4.7 percent of the industry’s annual revenues, or an additional $200 billion to $340 billion, according to McKinsey. On top of that impact, the use of Gen AI tools could also enhance customer satisfaction, improve decision-making and employee experience, and decrease risks through better monitoring of fraud and risk. 

Simply put, combining Gen AI with a platform technology can create an extremely powerful and unmatched experience. With Gen AI, banks can create and orchestrate digital workflows on a platform that foundationally helps them connect every system and digital process across their institution. This effort is critical as banks race to modernize their front- and back-office systems that have left them less able to compete with nimble tech startups. Such work becomes the foundation for an effective Gen AI strategy that, when combined with platform automation, will create a massive productivity catalyst that ripples across growth, cost-reduction and resilience.

Financial services companies are at significantly different points in their AI journeys. Surprisingly, some of the biggest players still lack this key strategy, while some smaller players are further along. This can’t start from the ground up: It has to be a team effort that works across – and integrates with – the industry’s notorious silos. The strategy can’t be disjointed and it should map up to offering better service, growing the business and improving productivity and innovation. Companies that try to have their Gen AI strategy live independently are less likely to succeed. 

There are four main ways financial services workflows may leverage Gen AI:  

Acquire and onboard customers' ability to grow the portfolio within compliance objectives. Financial services companies will be able to grow their portfolio within compliance objectives. The lengthy list of potential benefits includes summarization and generation by automating and personalizing document creation that requires only incremental information for new services or existing customers.

Regarding generation, imagine a world where they’ll be able to build a template for a new portfolio, loan or policy to speed customer acquisition by bypassing known information and only asking for net new or expired data. Meanwhile, they could benefit from customer segmentation and sentiment analysis to help identify profitable growth opportunities.

Provide frictionless servicing to see increased customer LTV at a reduced cost of service. They could be able to provide agents with a portfolio view of the customer that highlights recent and pending activity. Another benefit could be to automate after-interaction work including case coding, summarization and customer communication. A third potential scenario is a virtual agent that avoids costs and regulation for inquiries that may be resolved without cumbersome case creation./p>

Manage compliance and risk to maintain compliance with policies and regulations. A knowledge summarization could prove efficacy to auditors by demonstrating compliance and effectiveness relative to a list of customers, accounts or cases. Financial services companies may benefit from data analysis that interprets datasets and distills insights of incidents for improved operational risk management, while summarization will automate risk and control self-assessment reports to review incidents that occurred during the period, such as outages or risk events.

Modernize technology to reduce IT operating cost with modernization. Banks will be able to avoid disruption by proactively summarizing IT service issues and their likely causes and utilizing that data to drive preventative actions. They may be able to avoid complex coding by referencing documents, such as payment association rules, as the driver(s) of a playbook. Finally, summarization content could generate post-incident reviews and publish findings to regulators and third parties.

The time for banks to embrace Gen AI is here. Employees need not worry about losing their jobs - they should feel enabled to work smarter in ways that benefit them and the company - and the entire industry. By allowing Gen AI to imbue all aspects of the operation, financial services companies will benefit as they secure new customers with fast and personalized service while managing risk and compliance with modernized technology. Be a first mover.

Gregory Kanevski is ServiceNow’s Global Head of Banking.

Sacramento, California-based Golden 1 Credit Union has seen a 28% increase in approvals to protected classes of borrowers since it started using artificial intelligence for loan decisions, its CEO, Donna Bland, said in March at America’s Credit Unions’ Governmental Affairs Conference.

The credit union knew it was losing solid loans to other lenders due to faulty data, Bland said. So Golden 1 partnered with Zest AI to build its own customized credit scorecard.

Golden 1 is using data not only from the credit union’s membership in the algorithm but also from other Californians “who looked like our membership” to help grow the database, Bland said. 

Golden 1 launched the scorecard in December 2022 — first with credit cards and, then in 2023, with unsecured and auto loans. 

The results have been more powerful than the credit union even envisioned, Bland said. The AI-driven scorecard led to higher loan approvals overall but especially among protected classes, including women, Latinx and Black people, and members older than 62, the executive said. 

“What that means for us is that we’re no longer denying tens of thousands of people,” Bland said. “We originate over $6 billion in loans a year … and that not only means Californians have more access to equitable credit, it's also good for the credit union and our financial position.”

Oak Ridge, Tennessee-based ORNL Federal Credit Union also partnered with Zest AI to help the institution make better lending decisions, CEO Jenny Vipperman said.

For the $3.6 billion-asset credit union, “one size does not fit all” when it comes to rating potential borrowers.

“A national score was the best that we had when it was the best that we had,” Vipperman said. “But today, I’m looking for a score that can look at my fellow Tennesseans and saying ‘what they’re using to make a decision in California cannot possible make the most accurate decision in Tennessee.’”

Bland said more relevant information is critical for scoring potential borrowers and making better loan decisions. She said Golden 1’s membership is composed completely of Californians, but that demographic is different from the nation as a whole. 

“California financial-wise can be very different from other parts of the nation,” she said. “So relying on national and traditional scoring models can paint the wrong picture when the credit union is making loan decisions.”

At $21.1 billion in assets, Golden 1 is the sixth-largest credit union in the country but the biggest using AI for lending.

Jennifer Golbeck, director of the social intelligence lab at the University of Maryland, said financial institutions must be careful when deploying AI for credit decisions. She pointed to a situation in 2008, where the algorithm being used by Wells Fargo to make mortgage decisions was found to have racial biases.

Race was not a part of the data, but the algorithm picked up on the fact that decisions made by Wells employees led to certain populations being pushed into subprime mortgages where whites with the same financial profile were not.

"If humans are racially biased in the way that they’re handing out mortgages, the algorithm is going to learn to do that, too,” Golbeck said. “AI has this veneer of objectivity, but because what it does is replicate our decision-making, it replicates our biases as well.”

Morgan Stanley has tapped firm veteran Jeff McMillan to guide its artificial intelligence implementation as banks further explore uses for what some have deemed “the new electricity.”

The position, head of firmwide AI, is a new role at Morgan Stanley, though the bank — and McMillan, in particular — have been immersed in AI for some time. McMillan oversaw the development of AI @ Morgan Stanley Assistant, a virtual assistant for financial advisers and bankers powered by OpenAI, which launched in the fall.

McMillan’s appointment was announced in a March internal memo.

“In his new role, Jeff will coordinate across the Firm to ensure we have the appropriate AI strategy and governance in place,” wrote co-presidents Andy Saperstein and Dan Simkowitz. “In doing so, he will partner with the business units and infrastructure areas to identify and prioritize AI opportunities; help position the Firm within the flow of AI development across the industry and ensure that Morgan Stanley continues to be a well-respected innovator in AI.”

Users of the bank’s AI assistant reported its “meaningful incremental value,” McMillan said in January.

“Since rolling out the Assistant we have seen significant increases in accuracy, coupled with significant decreases in incomplete answers, when compared with our previous knowledge management process,” he said in a statement emailed to Banking Dive. “Further, the Assistant has reduced inquiry fulfillment times for Morgan Stanley’s Financial Advisor population, further driving efficiency and productivity.”

A Wall Street giant creating a role solely focused on AI is an ode to the technology’s growing importance in the banking sphere, with other banks sure to follow suit. AI has alternately been called “the new electricity,” “the new internet,” and “the new Industrial Revolution.”

In a study released in October during Money20/20, Zac Maufe, global head of regulated industries at Google Cloud, said generative AI “can represent a massive productivity and operational efficiency opportunity” and “help create hyper-personalized digital experiences that consumers are demanding,” especially as banking consumers become more comfortable with using the technology.

At Morgan Stanley, McMillan will report directly to Saperstein and Simkowitz and will collaborate on AI strategy execution with Mike Pizzi, head of U.S. banks and technology; Sid Visentini, head of firm strategy; and Katy Huberty, head of global research.

Huberty and McMillan will co-chair the firmwide AI steering group.

McMillan joined Morgan Stanley in 2009 and became head of investment platforms in the bank’s wealth management unit, according to his LinkedIn profile. He had served as the bank’s chief analytics and data officer since 2016. Before moving to Morgan Stanley, McMillan spent eight years as in Merrill Lynch’s research unit, according to LinkedIn.

JPMorgan Chase CEO Jamie Dimon stressed his commitment to cloud in his annual letter to shareholders in April.

“Getting our technology to the cloud — whether the public cloud or the private cloud — is essential to fully maximize all of our capabilities, including the power of our data,” he said.

The banking giant aims to have 75% of its data and 70% of its applications migrated to the cloud by the end of the year, Dimon said. Roughly 70% of its data and 50% of its applications now run in public or private cloud.

Cloud adoption is part of JPMorgan's broader prioritization of data to drive artificial intelligence adoption.

“In addition to making sure our data is high quality and easily accessible, we need to complete the migration of our analytics data estate to the public cloud,” Dimon said. “These new data platforms offer high-performance compute power, which will unlock our ability to use our data in ways that are hard to contemplate today.”

Dimon’s cloud enthusiasm was tied to the disruptive potential of generative AI and the data-hungry large language models that power the technology.

“We have large amounts of extraordinarily rich data that, together with AI, can fuel better insights and help us improve how we manage risk and serve our customers,” Dimon said.

JPMorgan Chase has more than 2,000 AI, machine learning and data science practitioners on its technology team and more than 400 AI use cases in production, he noted in the letter.

“We have been actively using predictive AI and ML for years,” Dimon said, noting the company is exploring additional use cases in customer service and operations, software engineering and employee productivity.

Banking stands to gain more from AI adoption than any other industry, according to an Accenture 2024 banking trends report. And cloud is crucial to scaling AI capabilities and unlocking the technology’s value, an Accenture executive said.

“When you have a business model that is fundamentally digital, the opportunity to apply the next generation of digital capabilities is unparalleled,” Michael Abbott, senior managing director and global banking lead at Accenture, told CIO Dive in January.

Cloud is the backbone of the digital enterprise and an area of heavy investment for JPMorgan Chase. The company poured $2 billion into the construction of four U.S.-based private-cloud facilities announced in 2020 and now has 32 global data centers, Dimon said.

The bank also invested in data operations, appointing Teresa Heitsenrether to the newly created chief data and analytics officer position in June 2023.

“Elevating this new role to the Operating Committee level — reporting directly to [President and COO] Daniel Pinto and me — reflects how critical this function will be going forward and how seriously we expect AI to influence our business,” Dimon said in his letter. “This will embed data and analytics into our decision making at every level of the company.”

Banks are balancing ongoing spending on cloud and cybersecurity technologies against an eagerness to invest in artificial intelligence, according to a Broadridge survey of 500 financial services C-suite and senior executives in 18 countries published in March.

The vast majority of respondents are prioritizing data management, cybersecurity and cloud platforms in day-to-day investments. They also intend to maintain or increase spend on emerging technologies, including generative AI, quantum computing and blockchain over the next two years.

Digital transformation has helped lay the groundwork for AI adoption in banking, which is poised to benefit from the technology more than any other industry, according to Accenture. More than two-thirds of Broadridge respondents said they’ve made meaningful progress in modernizing their core IT platform.

Cloud providers inundated enterprises with a steady stream of model options in 2023, and the tap has yet to run dry. Nvidia unveiled a portfolio of dozens of enterprise-grade, cloud-based microservices in March aimed at helping companies tune models with proprietary data.

But transformational solutions require a clean data estate, a robust tech stack and heightened security measures.

“With pre-trained models, companies are finding that it is possible to make compelling progress on previously stalled use cases and pilots,” Joseph Lo, head of enterprise platforms at Broadridge, said in an email. “However, it is also clear that companies that have made big moves with generative AI have differentiated themselves with the data that is proprietary to them.” 

Investments in operationalizing generative AI cut across the business, from front-office customer service functions and document summarization capabilities, to beefing up data management systems, the survey found.

“Most companies don't have any shortage of data,” Lo said. “But the sophistication of being able to access it at the velocity that is required to hydrate these models and the use cases they enable has been the challenge.” 

Security is a problem, too.

Cyber concerns continue to stall digital transformation, prompting financial firms to boost investment in this area by an average of 28% in the next two years. Security was also the top capability executives expect from third-party vendors, the survey found.

“Data readiness — the ability to access proprietary, clean data at sufficient velocity with the right levels of security, is table stakes,” Lo said. "Companies must ensure they have robust AI governance policies and controls, inclusive of cyber resilience, privacy, and data security, in place.”

The rapid growth of generative artificial intelligence has created a looming cybersecurity risk for the banking sector, the Federal Reserve’s vice chair for supervision, Michael Barr, warned in October.

As use cases for generative AI continue to evolve, “an arms race between those who are using generative AI to attack and those who are trying to use generative AI to block the attack” will ensue in the financial sector, Barr said during a live-streamed cyber risk conference held by the Federal Reserve Bank of Cleveland.

“Technology is always evolving, and so we have to keep up with the next threat. Generative AI could be a really important new concept for the economy, for finance, for really many aspects of economic life. And as with any new technology, it will have risks,” he said.

Barr said it is critical for financial institutions to make investments in generative AI to safeguard against cyberattacks. 

“There's a real risk that we have a cyber arms race using generative AI with defenders and attackers in a constant struggle,” Barr said. “So we do need to make sure that we are, and banks are, investing in the kind of technology that is useful, not only today, but in the near future.”

Barr also highlighted third-party relationships, emphasizing that banks of all sizes have a responsibility to ensure their vendors are properly managing cyber risk. 

“The banks need to treat the third-party risk management as if it were their own risk inside the institution, because legally it is, and from a supervisory perspective, it's required. It's just essential in terms of the safety and soundness of the financial system,” he said.

While it is mostly smaller institutions that rely heavily on third-party service providers, Barr noted that large institutions moving to cloud-based solutions also need to be vigilant. 

“It's absolutely essential that as they do that, banks are ensuring that their relationships with third parties, not only are not hurting them, but are actively helping them to manage their cyber risk,” he said. “We really need to have resiliency in critical service providers, and the banks have an absolutely central role in making sure that that's the case.”

Ripple effects

Due to the tightly interconnected nature of the U.S. financial system, a data breach at one firm can have devastating ripple effects across the financial services sector, even if that firm is a smaller bank, Barr said.

“If there's a problem at a small bank, it can then infect the rest of the financial system,” Barr said. “We're not any more safe if it's a small bank than a big bank, because that problem can then grow into the rest of the system. We have to worry about even the weakest link in the system.”

A cyberattack at a financial institution could disrupt a firm’s payment system and liquidity provision channels, Barr said. 

“If a bank has to cut itself off in order to protect the rest of the system from a cyberattack, that cutting off has ripple effects that could cause bank failures,” he said.

Barr emphasized the importance of having firms engage in internal and external testing, adding banks need to find creative ways to ensure they have the ability to recover and resume operations with the least amount of client disruption.

Barr’s comments come as Consumer Financial Protection Bureau Director Rohit Chopra has also warned of the risks of generative AI, albeit from a consumer perspective. 

Chopra, in April 2023, said the bureau is closely monitoring how generative AI and ChatGPT-like technologies, when used by banks, could undermine or create risks in customer care.

“We’re doing some work right now on how [generative AI] might undermine or create risks in customized customer care, to the extent that biases are introduced, or frankly, even the wrong information,” he said.

Meanwhile, an interagency initiative launched by the Justice Department, Federal Trade Commission and Equal Employment Opportunity Commission in April aims to crack down on “unchecked AI” in lending, housing and employment.

President Joe Biden signed an executive order in October aimed at regulating the development of artificial intelligence. 

The order sets “new standards for AI safety and security, protects Americans’ privacy, advances equity and civil rights, stands up for consumers and workers, promotes innovation and competition, advances American leadership around the world, and more,” the Biden administration said in a fact sheet.

The order comes as the banking sector, like most others, has already been dipping its toes into the nascent space, exploring the many internal and external use cases for AI and machine learning technology. 

Unsurprisingly, the topic came up just about everywhere at October’s Money20/20 conference in Las Vegas, both on the stage and during attendees’ one-on-one discussions with Banking Dive reporters.

To give a feel for how banking, fintech and payments leaders are approaching the technology, Banking Dive compiled a roundup of executive insights from the conference.

Overheard at Money20/20 ...

 

While Dave, a neobank and personal finance app, has been using artificial intelligence to underwrite its products since the company’s inception in 2017, the fintech’s latest use of generative AI is producing promising results on the platform’s customer service side, said CEO Jason Wilk.

The fintech in December launched DaveGPT, a gen AI-driven chatbot that can respond to customer inquiries in real-time. 

The tool, which replaces its DaveBot, is generating an 89% resolution rate, Wilk said.

“It’s pretty impressive, based on what we used to have with our DaveBot, which was significantly lower,” he said.

DaveBot’s “rules-based model” would generate canned responses, compared to DaveGPT’s “more human, large language model approach,” Wilk said. 

“People really are getting much more to the bottom of what they're looking for,” he said. “On the consumer side, we know that people want real-time responses these days, especially millennials and Gen Z, which is the majority of our audience at this point. They're not wanting to talk to somebody on the phone, they want real-time chat responses if they can get it. And AI is certainly the best way to get a most human-like response with the best accuracy.”

Dave used an existing partnership with Aisera, which leverages OpenAI’s capabilities, to access large language models, adjust the content and then come back with the relevant responses, Wilk said, adding that he envisions a future iteration of the product that involves integrating customers’ financial data to offer personalized advice.

“The AI can recommend solutions to your problems, such as, ‘How do I save more money this month?’” he said.

With the rollout of DaveGPT, Dave becomes the latest neobank to fold AI into its offerings. 

Neobank MoneyLion, in May 2023, detailed plans to launch a search feature powered by ChatGPT to allow consumers to search for transactions, offers and products on the platform’s finance marketplace.

AI helps fintechs like Dave serve more customers without staffing up, Wilk said. The technology allows Dave to serve its 9 million customers with only 300 employees, Wilk said.

“We have a lot of operating leverage compared to incumbent banks,” he said. “AI is another lever we have to be able to do more while not growing our headcount in any kind of major way. So that's a really important way that we can scale. We don't need 1,000 more people for customer support. If anything, we can keep our support team the way it is and this allows us to 10x our capabilities.”

Beyond using AI in its customer support, Wilk said the technology has become a critical part of the firm’s operations. 

“DaveGPT is just our customer-facing support program, but we're using AI all the way across the company,” he said, adding the fintech uses AI to help underwrite its cash-advance product.

“It's really been a critical way for us to improve credit access by giving people access to higher limits, while also keeping our loss rates at a record low,” he said. “The AI can look at thousands of pieces of data from our cash flow-based underwriting that enables us to make better decisions, which again, helps us provide a better experience for credit access with better limits and reduces our loss rates.”

Wilk said he expects gen AI will continue to grow in popularity in the banking and fintech space, both as an underwriting tool and as a means to improve customer satisfaction.

“I think that first and foremost, AI is going to be a game changer for credit,” he said. “And then the customer support stuff will be sort of table stakes which everyone's going to need to keep costs low.”

Bunq in December became the latest neobank to fold generative artificial intelligence into its platform.

The Amsterdam-based fintech said its generative AI tool, called Finn, is available to all 11 million of its users.

“Finn will wow you,” Ali Niknam, Bunq’s founder and CEO, said in a statement. “Years of AI innovation, coupled with laser focus on our users, allowed us to completely transform banking as you know it. Seeing generative AI make life so much easier for our users is incredibly exciting.”

Finn replaces the search function on Bunq’s app, the firm said in a release, allowing consumers to plan their finances, budget, navigate the platform and find transactions.

Users can ask questions or seek advice about their bank account and spending habits in Finn’s “chat-style text box” the company said. 

The platform can tell users their monthly grocery spend or how much they spent in a year at specific retailers like Amazon, the company said. 

“It can even combine data to answer questions that go beyond transactions, such as ‘What was that Indian place I went to with a friend in London?’, or ‘How much did I spend at the cafe near Central Park last Saturday?’” the company said.

With the launch of Finn, Bunq joins several other neobanks looking to leverage generative AI to improve the consumer experience.

Neobank MoneyLion, in May 2023, said it plans to launch a search feature powered by ChatGPT that would allow consumers to search for transactions, offers and products on the platform’s finance marketplace.

Dave, a neobank and personal finance app, in December launched DaveGPT, a generative AI-driven chatbot that can respond to customer inquiries in real time.