At America’s oldest bank, 134 new employees don’t sleep or call in sick. They don’t even have names.

They are what Bank of New York calls its “digital employees.” They work side by side with humans. They play unique roles and are evaluated based on their performance. Some of their work last year was done by people.

“Digital employees work 24/7, which is obviously very different from our human counterparts,” said Rachel Lewis, who oversees nine digital employees in addition to thousands of employees as head of payments at Bank of New York. “It really focuses on very specific repetitive tasks and allows our employees to take on roles that are more human, more intense, more interesting.”

Bank of New York has 48,100 employees, down from about 53,400 in 2023 Recent earnings reports. Last month, on the company’s fourth-quarter analyst call, Chief Financial Officer Dermot McDonogh was asked what 134 digital employees would mean for the company’s cost savings.

“Our headcount is down slightly, but that doesn’t have anything to do with AI yet,” McDonough said. “We talk internally about AI being unlocking capabilities. We don’t think about it in a narrow definition of efficiency. It’s all about growing with customers, growing revenue and optimizing the potential of our employees.”

Across Wall Street, analysts and investors are starting to ask more questions about how the industry’s spending on artificial intelligence will translate into greater efficiency and greater returns. In 2025, Bank of New York will spend $3.8 billion on technology, accounting for about 19% of its revenue. That’s the highest ratio among large bank peers, according to data compiled by CNBC.

“It’s an AI arms race. Banks are part of it,” said Wells Fargo analyst Mike Mayo. “But you can’t define success by who spends the most.” You can define success by who achieves the best results. “

“In general, there’s a lot of ‘spray and pray’ when it comes to technology spending,” he said.

However, Bank of New York is considered one of the companies that could benefit the most from artificial intelligence. Goldman Sachs’ research team screened the Russell 1000 for potential productivity gains based on the labor cost and wage risks of AI automation. The firm ranked Bank of New York at the top of the list and said the bank’s earnings per share could grow 19%.

But in several conversations CNBC had with New York bank executives, they were adamant that the massive investment in technology would not come at the expense of employees.

“I wouldn’t think so,” said Michelle O’Reilly, Bank of New York’s global head of talent. “I prefer to think of it as unlocking productivity—enabling all employees to be productive.”

The company is also upskilling its employees as it develops more digital employees. Shortly after ChatGPT was released in late 2022, the Bank of New York set up an AI Hub.

“That’s when we really doubled down on our efforts and realized this was going to be transformative for the bank,” said Leigh-Ann Russell, chief information officer and global head of engineering at Bank of New York. “Our biggest focus initially was support — providing some training to every employee in the bank.”

Bank of New York has built a platform called Eliza that introduces a variety of open source, commercially available models that integrate with the company’s internal data and compliance. Nearly all Bank of New York employees completed Eliza’s 10-hour training, and thousands more took it a step further with a multi-day AI bootcamp that helps non-engineers find creative ways to automate parts of their jobs.

The name “Eliza” commemorates Elizabeth Schuyler Hamilton, the bank’s founder and wife of Alexander Hamilton, the first U.S. Secretary of the Treasury.

“The democratization of this technology is one of the sweet spots where we feel like we’ve been successful so far,” Russell said. “I juxtapose the raw history of this amazing 241-year-old institution with being at the forefront of artificial intelligence, and I think it’s just a lovely reminder of centuries of technology.”