Marco Argenti, chief information officer at Goldman Sachs, is leading one of Wall Street’s most aggressive integrations of A.I. He has made a name for himself as an early adoptor of A.I. in finance through initiatives like the GS AI Assistant platform, which is offered to Goldman Sachs employees for tasks such as coding and translation, and last year’s pilot of A.I. software engineer Devin, made by Cognition Labs. More recently, the investment bank has been collaborating with Anthropic, using its Claude model primarily in its accounting and compliance departments, Argenti said in an interview with CNBC published today (Feb. 6).
The goal is to speed up tasks that involve massive amounts of data without investing in more manpower. “Think of it as a digital co-worker for many of the professions within the firm that are scaled, complex and very process-intensive,” Argenti said.
Argenti spent much of his career in the tech and cloud computing industries before joining Goldman Sachs in 2019. He previously served as vice president of technology at Amazon Web Services, overseeing serverless computing and virtual reality. Earlier in his career, he led developer experiences at Nokia.
Anthropic is known for its A.I. coding assistant, which is widely used by engineers. Goldman Sachs quickly realized that the traits that make a good coder—such as applying logic and working with large volumes of complex data—could be applied to tasks across accounting and compliance, Argenti said. Outside those departments, Claude agents could also be used for employee surveillance and creating investment banking pitchbooks for clients, he revealed.
Goldman Sachs and Anthropic did not respond to requests from Observer to comment on those efforts.
A collaboration with Goldman Sachs is the latest win for Anthropic, which has positioned itself as an enterprise-focused A.I. company. Earlier this week, the startup’s release of coworking software with various industry plug-ins triggered a panic selloff in enterprise software stocks, as investors worried such tools could make existing products obsolete.
Other Wall Street giants are also embracing A.I. agents. JPMorgan Chase currently has more than 500 A.I. use cases, ranging from customer service to idea generation and marketing, and draws upon models from both Anthropic and OpenAI to power its internal LLM Suite program. Morgan Stanley was an early client of OpenAI, using its tech to distill meeting notes, aid financial research and boost coding productivity.
A.I.’s use in financial services has grown each year since 2022, according to a recent Nvidia survey, in which 100 percent of industry professionals said A.I. spending will either stay the same or increase in 2026. A.I. agents, in particular, are being used or assessed by 42 percent of respondents. Top workflows include knowledge management and retrieval, internal process optimization and customer support automation.
Such widespread adoption will inevitably lead to industry-wide labor shifts. A.I. leaders and studies alike have warned that the technology could reshape or eliminate entry-level white-collar roles. It’s unclear how the use of A.I. would affect Goldman Sachs’ employees. But Argenti conceded that A.I.’s advancements could eliminate the need for third-party providers.
