How BNY is Banking on Innovation in AI and Digital Assets
By Steve Cocheo, Senior Executive Editor at The Financial Brand
Simple Subscribe
Subscribe Now!
Executive Summary
- BNY CEO Robin Vince says banks must keep an eye on innovation risk but also remember that failing to innovate is also risky. Two key focuses are AI and digital assets.
- The bank’s enterprise AI system, called Eliza, is already in its second generation and almost every BNY staffer has been trained on its agentic tools.
- BNY got in early on digital assets, but as an enabler/partner for other companies, with less interest in issuing its own stablecoins.
How does a 241-year-old megabank innovate?
Boldly but carefully, says Robin Vince, CEO of The Bank of New York Mellon, discussing its stance on digital assets and artificial intelligence.
“All innovation has risk associated with it,” according to Vince, “but to not innovate is very risky too.”
His case in point: Kodak discovered the essentials of digital photography. But the company’s leaders believed that it would be disruptive to its core business, traditional film-based photography. So, Kodak buried it and let others pioneer the way we take pictures today.
That said, a regulated bank “has to innovate in a smart, sensible way,” Vince told listeners at the recent annual conference of The Clearing House.
“You don’t want your people to go absolutely crazy,” so guardrails have to be in place, said Vince. To this end, BNY has line of business teams working on innovative ideas in cooperation with representatives of risk management.
“The best type of product development inherently includes a risk management element,” said Vince.
Want to read more like this? Check out Narmi’s content portal on The Financial Brand: Be Where Banking is Going
Here’s an example: As part of the company’s major push into AI, it has developed a cadre of over 100 “digital employees.” These are not humans who work on digital tasks. They are AI agents — software developed to perform a given set of tasks. Many of them actually possess bank user IDs, email accounts and access to the systems necessary for their jobs — just like human employees. The digital agents receive training from human staffers before they go live.
One such job is poring through “multimillions of lines of code” for vulnerabilities — an essential task that few humans want. Digital employees — the agents — assigned to this find the weak points. Then they write fresh code to fix the weakness. Then, having learned and devised an approach to the problem, they promote the change into the bank’s ongoing code cleanup process.
However, adoption isn’t happening in isolation. “It’s reviewed by a human engineer,” said Vince.
Read more: Is AI Learning the Job Faster Than Banks Can Redefine It?
Growing the Bank’s Own Multi-Agentic AI
A couple of years ago, Vince recounted, the bank made a bet on multi-agentic AI. Broadly, this means using multiple artificial intelligence agents that act together to accomplish tasks.
Having made that choice, he continued, the bank realized that many of the multi-agentic systems available worked well inside their own “walled gardens,” but that there was no way to tie in agents developed on other platforms.
“We wanted to be able to take full advantage of other companies’ agents, but we didn’t want to be confined to just using one set of agents, or to have a series of different walled gardens that didn’t talk to each other,” said Vince. The bank wants to be able to continuously add “best in class” agents to the selection offered to staff.
The solution: Building the bank’s own horizontal multi-agentic AI system, one capable of plugging in multiple agents from many sources. The system was dubbed “Eliza.” It’s the namesake of Eliza Hamilton, wife of Alexander Hamilton, who helped found the bank. In technical terms, Eliza is an enterprise AI platform. Eliza 2.0 launched in September.
Vince says that 99% of the bank’s employees have had the initial training in how to use Eliza. Overall, 20,000 of the bank’s 50,000 staffers have used the technology at some point to experiment with building agents for one task or another.
As a result, “we have a bunch of power users who aren’t actually in our engineering group,” said Vince.
To put some further definition on “a bunch”: An October white paper on the bank’s site says that most of the bank’s AI builders now come from outside of its engineering teams. It considers Eliza an approach to “democratizing” access to AI tools. The white paper details how BNY approaches training for AI usage. Methods include self-guided elearning, intensive boot camps, hackathons, bank AI communities, and opportunities for hands-on learning.
Read more:
Will BNY Cut Staff as AI Takes Hold?
Moderator Leo Schwartz, senior reporter at Fortune, noted that two major Wall Street financial organizations have indicated that AI adoption is influencing staff levels. Goldman Sachs has been laying people off, and JPMorgan Chase has said AI will change its approach to hiring. What is BNY’s philosophy? he asked.
Vince said that the bank sees AI as a means of “workforce augmentation. We are a large firm, a scale provider, but we aren’t mega size.” Management wants to contain its expense base, and a key part of the AI push is to gain the ability to grow without scaling up expenses.
However, “we’re not primarily looking at AI as a cost-cutting opportunity,” said Vince. “Our first instinct is to deploy AI as a superpower for everyone in the company — make AI pervasive everywhere, make it really a ubiquitous capability.”
Read more: How the Agentic AI Revolution is Transforming Operations at 70% of Banks
Digital Assets: Opportunities to Speed Payments and Commerce
When asked if he saw the bank ever issuing its own stablecoin, Vince said he thinks the bank’s most appropriate role in digital assets is facilitating other companies’ efforts.
“We’re in the business of enabling other people’s business models first and foremost,” said Vince, though he wouldn’t rule out issuing a stablecoin. However, he said, “I don’t love the idea of having to launch our own stablecoin. I would much rather be in the business of providing capabilities to enable other stablecoins to prosper. That’s the DNA of our company.”
Vince said that he could see providing white label technology for corporations that wanted to issue their own stablecoin, possibly even just for internal use.
In its third quarter earnings briefing, Vince noted that BNY is playing a long-term game in crypto, having begun to provide related services three years ago. For example, it was an early big bank entrant in digital asset custody. In his fireside chat Vince noted that that early involvement has brought a good deal of traditional business to BNY from digital asset companies that saw the bank’s efforts as indicative of having an open mind.
“It’s actually been a great source of business for us,” said Vince.
Read more: Will the GENIUS Act Revolutionize Banking and Payments? Absolutely. Here’s How
Blockchain and Digital Assets: Greasing the Wheels of Business
Vince sees blockchain technology as improving flows of commerce in multiple ways.
Moving money faster and with certainty is one of them. He says this need might not have been the same if the U.S. were further along with instant payments techniques.
“It’s not too late, but we should really get on that if we want to make sure that traditional money movements continue to dominate the financial system in the next ten years,” said Vince.
Another key reason for exploring blockchain payments is the clunkiness traditional rails present for some purposes. Vince said that settling a loan in traditional ways, for example, can be “a messy business.”
Finally, he said, “if you’re going to operate in more of an on-chain world, you need an on-chain currency.” Cryptocurrencies like Bitcoin and Ethereum experience considerable price volatility and therefore are problematic when someone is trying to pay for something on-chain with a currency lacking a stable value.
By definition, stablecoins, backed under the GENIUS Act by assets such as U.S. Treasury securities. So Vince said using them, rather than more volatile cryptocurrencies, makes sense. (GENIUS stands for “Guiding and Establishing National Innovation for U.S. Stablecoins.”)
He added that his message to traditional banking players is they should not only be innovating with digital assets, but also continuing to innovate in traditional payments — “so we don’t get completely overrun by a new digital asset ecosystem.”
Meanwhile, added Vince, regarding BNY, “being a bridge between digital asset rails and traditional rails is a good place for us to be.”
Read this next: How to Blunt the Threat of Smart Contracts and Stablecoins to B2B Payments
