Goldman Sachs has been working with artificial intelligence startup Anthropic for the past six months to co-develop autonomous “AI agents” intended to automate trade and transaction accounting and client vetting and onboarding, CNBC reported. 

The investment banking giant has embedded Anthropic engineers alongside its own teams as it builds agents based on Anthropic’s Claude model, Goldman CIO Marco Argenti revealed. The initiative is still in the early stages but is aimed at sharply reducing the time required to complete process-heavy functions that are essential to the bank’s operations, according to the report. 

“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,” Argenti told CNBC, adding that he expects the agents to launch soon. 

Goldman executives have been surprised by Claude’s effectiveness beyond software development, particularly in accounting and compliance work that requires digesting large volumes of data and documents while applying rules and judgment. They now believe that AI agents could accelerate customer onboarding and other tasks, the news agency said. 

The plan comes amid a multiyear plan to reorganize the investment bank around generative AI, according to CNBC. 

While Goldman employs thousands of people in the compliance and accounting functions where agents are expected to operate, Argenti told the news agency that it is “premature” to assume job losses. He said the technology may allow the bank to reduce reliance on third-party providers over time, but in the near-term the institution is focused on building capacity to move faster, improve the client experience, and potentially drive more business.

Read more at CNBC