Goldman Sachs is deploying AI agents, developed in partnership with artificial intelligence startup Anthropic, to automate a growing number of roles within the financial institution. Marco Argenti, Goldman’s chief information officer, revealed that for the past six months, the bank has been collaborating with embedded Anthropic engineers to co-develop autonomous agents for at least two key areas: trade and transaction accounting, and client vetting and onboarding.
These agents, built upon Anthropic’s Claude model, are in the “early stages” of development and are expected to significantly reduce the time required for these essential functions. Argenti anticipates a launch “soon,” though he stopped short of providing a specific timeline. He described these AI agents as “digital co-workers for many of the professions within the firm that are scaled, are complex and very process intensive.”
This initiative aligns with Goldman Sachs CEO David Solomon’s October announcement of a multiyear plan to reorganize the bank around generative AI. Even as Goldman Sachs experiences robust revenue from trading and advisory activities, the firm intends to “constrain headcount growth” through this strategic overhaul. The move by Goldman Sachs comes amidst a backdrop of intense investor scrutiny on the AI sector, following recent model updates from Anthropic, co-founded by a former OpenAI executive, which have reportedly triggered significant sell-offs in software companies and their lenders as the market assesses the emerging leaders and laggards in the AI landscape.
Last year, Goldman Sachs began experimenting with an autonomous AI coder, a tool now widely accessible to the bank’s engineers. However, Argenti noted that the firm soon discovered Anthropic’s AI model’s capabilities extended beyond coding. “Claude is really good at coding,” Argenti remarked, posing the question, “Is that because coding is kind of special, or is it about the model’s ability to reason through complex problems, step-by-step, applying logic?”
The bank was reportedly “surprised” by Claude’s proficiency in tasks beyond coding, particularly in complex domains like accounting and compliance. These areas demand the parsing of vast amounts of data and documentation, coupled with the application of stringent rules and judgment. Consequently, the prevailing sentiment at Goldman Sachs is that “there are these other areas of the firm where we could expect the same level of automation and the same level of results that we’re seeing on the coding side.”
The anticipated outcome is a streamlined client onboarding process and faster resolution of trade reconciliation and other accounting-related issues, facilitated by these developing AI agents. Looking ahead, Goldman Sachs is exploring the development of AI agents for tasks such as employee surveillance and the creation of investment banking pitchbooks.
While thousands of employees are currently engaged in compliance and accounting functions where these AI agents will be deployed, Argenti stressed that it is “premature” to anticipate widespread job losses. He did, however, suggest that as AI technology matures, Goldman Sachs might reduce its reliance on certain third-party providers. “It’s always a trade-off,” Argenti concluded. “Our philosophy right now is that we’re injecting capacity, which in most cases will allow us to do things faster, which translates to a better client experience and more business.”
Original article, Author: Tobias. If you wish to reprint this article, please indicate the source:https://aicnbc.com/17169.html
