BNY, Goldman double down on tech with 'digital employees' and AI agents

BNY and Goldman Sachs are pouring billions of dollars into artificial intelligence and automation, creating “digital employees” and AI agents that are starting to handle core operational work.

At BNY, 134 software-based “digital employees” now work alongside the bank’s roughly 48,100 human staff, taking on functions that were done by people just a year ago.

“The digital employee works 24/7, which is obviously very different to our human counterparts,” Rachel Lewis, head of payment operations at BNY, told CNBC. “It’s really focused on very specific repetitive tasks that allow our human employees to do much more human, intense, interesting-type roles.”

BNY reportedly spent $3.8 billion on technology in 2025, about 19% of its revenue – the highest share among its large-bank peers. While the bank’s headcount has drifted lower from about 53,400 in 2023, finance chief Dermot McDonogh has framed AI as a way to expand capacity rather than as a straightforward cost-cutting tool. He told analysts the firm views AI as a way to grow with clients, increase revenue and “optimize the potential” of employees.

Wall Street analysts are watching closely to see whether that spending translates into better returns. “There’s an AI arms race. The banks are part of that,” Wells Fargo analyst Mike Mayo said, adding that success will be judged not by capital spending – a hotly scrutinized question across the corporate world – but “who has the best results.”

External research suggests BNY could be one of the winners. Goldman Sachs analysts ranked the company among the firms most likely to see productivity gains from AI, estimating it could achieve about a 19% boost to earnings per share as automation spreads through labor-intensive functions.

BNY executives, however, have been careful to emphasize that their AI build-out is designed to augment, not replace, people. Along that seam, it launched an internal AI Hub and built a platform known as Eliza shortly after ChatGPT’s 2022 debut, connecting a mix of open-source and commercial models with the bank’s data and compliance systems. Nearly all employees have completed a 10-hour Eliza training, and thousands have joined multi-day AI bootcamps aimed at helping non-engineers automate parts of their jobs. 

Goldman Sachs is taking a parallel path with its work on AI agents. For the past six months, the bank has reportedly been collaborating with Anthropic engineers to develop agents based on the Claude model for two high-stakes areas: accounting for trades and transactions, and client vetting and onboarding.

“Think of it as a digital co-worker for many of the professions within the firm that are scaled, are complex and very process intensive,” Marco Argenti, Goldman’s chief information officer, told CNBC separately. The goal is to compress the time it takes to reconcile trades or onboard clients, with the expectation that automation will later move into areas like employee surveillance or investment banking pitchbooks.

Goldman has already rolled out an autonomous AI coder called Devin to its engineering teams and found that Claude’s strengths extend beyond software development. Argenti said the firm was “surprised” by how well the model handled accounting and compliance work that blends document-heavy workflows with the need for rules-based judgment.

Even as Goldman CEO David Solomon has talked about constraining headcount growth as part of a multiyear reorganization around generative AI, Argenti called it “premature” to assume direct job cuts in compliance and accounting. Over time, he said, the bank may rely less on third-party vendors as in-house AI tools mature.

These moves come against a backdrop of rising expectations around technology in the banking channel. Recent research by Cerulli Associates found 80% of banking advisors say a firm’s technology stack influences whether they stay or move, as digital tools shift from a back-office concern to a core part of the value proposition. Tellingly, advisors report AI usage at 29% in retail bank and bank trust channels, compared with 56% in private banks in 2025.

“The largest overall challenge facing bank wealth management programs is their ability to hire, train, and retain productive & profitable advisors, and 80% of these advisors are saying they seriously consider technology as a part of a firm's offering when deciding where to conduct their business,” Matt Zampariolo, research analyst at Cerulli, told InvestmentNews.