Goldman Sachs Partners with Anthropic to Automate Banking Tasks
On February 8, 2026, Goldman Sachs announced a collaboration with AI startup Anthropic to develop “AI agents” aimed at automating internal banking tasks. These tasks include essential functions like trade and transaction accounting and client due diligence, which is the process banks utilize to verify their customers.
CNBC reported that this strategic shift is notable as it moves beyond simple chat assistants to software capable of independently handling multi-step tasks. Historically, banks have approached such automation cautiously due to the strict regulatory demands of the industry.
Current Applications and Future Potential
The AI agents are currently being trialed for tasks such as transaction reconciliation, trade accounting, client vetting, and onboarding. According to PYMNTS, these tasks have resisted automation for decades, making this development particularly significant.
Goldman Sachs’ Chief Information Officer, Marco Argentini, stated that the firm is “in the early stages” of this initiative and expects to roll out these agents “soon.” While he did not specify a timeline, he labeled concerns about job cuts as “premature.” He did acknowledge that the bank might reduce reliance on certain third-party providers as the technology evolves.
Transforming Banking Roles
Argenti emphasizes that the AI agents should be viewed as digital co-workers rather than merely a technological spectacle. He stated, “Picture it as a digital co-worker handling roles within the firm that are scaled, complex, and heavily process-driven.” Future applications could also include employee monitoring and crafting investment-banking pitchbooks.
Industry Context and Competitive Landscape
The move towards AI by Goldman Sachs is not isolated. Other financial institutions are also exploring automation. For instance, in September 2025, Citib announced an upgrade to its proprietary Citi Stylus Workspaces, incorporating Agentic AI. Similarly, Morgan Stanley rolled out “Debrief,” a tool powered by OpenAI for generating meeting notes. JPMorgan has also revealed plans for an LLM Suite that will include AI agents capable of executing multi-step processes.
Regulatory Challenges
Despite the promising capabilities of AI, the risks associated with these technologies are significant. In December, the Financial Conduct Authority in Britain flagged “agentic AI” as posing new dangers due to its ability to act rapidly, which could exacerbate governance and stability issues. The initial challenges Goldman faces will focus on the regulatory concerns, including audit trails, controls, and ensuring software reliability, especially under complex conditions.
As these AI agents are developed, the financial sector will need to address not only the potential efficiencies but also the inherent risks associated with automation in highly regulated environments.
