WASHINGTON — In another indication of how AI is further penetrating financial institution operations, Goldman Sachs is deploying autonomous artificial intelligence agents built with Anthropic’s Claude model to automate core accounting, compliance and operational finance functions, according to a new report.
The initiative underscores the growing use of so-called agentic AI across corporate finance, as firms test productivity gains while navigating regulatory risk and internal controls, CNBC sai din its analysis.
Goldman Sachs Chief Information Officer Marco Argenti told CNBC the firm spent six months embedding engineers from Anthropic inside its technology teams to co-develop AI agents capable of handling complex, rules-based financial tasks. Those agents are being tested on transaction reconciliation, trade accounting, and client vetting and onboarding — functions long considered difficult to automate because they require processing large volumes of data under strict regulatory frameworks.
Earlier Experiments
The CNBC report noted that Goldman’s push follows earlier experiments with AI coding assistants that convinced engineers the technology’s reasoning capabilities were sufficient for more advanced financial use cases, Argenti said. Chief Executive David Solomon has previously described generative AI as central to a multiyear strategy aimed at limiting headcount growth while accelerating internal workflows.
Internally, the firm is positioning the tools as “digital colleagues” that augment human workers rather than replace them, CNBC said.
“Still, market reactions suggest broader tensions,” the report stated. “A recent sell-off in technology and financial stocks following the release of a new automation tool by Anthropic highlighted investor concerns that AI could disrupt legacy software vendors and reduce labor needs across industries.”
‘Wider Trend’
Goldman’s effort reflects a wider trend among large financial institutions. Citigroup has rolled out Stylus Workspaces, an internal platform designed to automate complex, multi-step finance tasks across applications while keeping sensitive data and compliance logic in-house.
Adoption remains cautious. Surveys cited by CNBC show CFOs are most comfortable using AI in structured, rules-based areas such as cash flow monitoring and compliance, while maintaining human oversight for higher-risk decisions. Still, interest is rising, with a growing share of finance leaders exploring agentic AI for planning, reporting and cost management as the technology moves from pilot programs into live production.
