In an effort to modernize core banking functions and increase operational efficiency, Goldman Sachs Group Inc. has initiated a partnership with the artificial intelligence startup Anthropic. The bank is collaborating closely with embedded engineers from Anthropic to develop AI-driven agents that will automate critical banking activities. This joint venture has been underway for six months, focusing on deploying artificial intelligence within the bank’s daily processes.
During a discussion with CNBC, Marco Argenti, the chief information officer at Goldman Sachs, elaborated that the collaboration primarily targets two specific banking functions. The first area is trade and transaction accounting, a domain that involves managing and reconciling intricate financial records. The second focus area centers on client vetting and onboarding procedures, which generally require thorough regulatory checks and compliance steps before integrating new clients into the system.
As the project remains under active development, the partnership has not disclosed a definitive timeline for the official rollout. However, Argenti noted that the deployment is anticipated to happen in the near future. He emphasized the nature of the AI solution being co-developed, describing it as a “digital co-worker.” This AI agent is intended to handle roles characterized by scale, complexity, and process intensity, thereby supporting human staff in managing demanding tasks.
Anthropic, the AI provider collaborating with Goldman Sachs, recently upgraded its offerings with the launch of Claude Opus 4.6. This iteration features improved capabilities in planning, coding, and debugging, which enable more sophisticated interaction with complex tasks. The company projects significant revenue growth, estimating it will achieve $18 billion by 2026, with a strong enterprise customer base contributing 85% of their income. Notable clients include global financial institutions such as JPMorgan Chase, which has also integrated Anthropic’s technology solutions.
The AI technology has garnered recognition beyond the financial sector, with NVIDIA’s CEO Jensen Huang praising Claude’s ability to execute coding and reasoning functions during the World Economic Forum in Davos. This endorsement highlights the advanced reasoning capabilities of the AI platform, indicating robust potential for diverse professional applications.
Goldman Sachs' strategic leader, CEO David Solomon, has been vocal about the company’s commitment to generative AI. Announced in October, the bank’s multi-year internal transformation plan revolves around integrating AI-driven technologies while maintaining disciplined control over headcount growth. Prior to the current collaboration, the bank experimented with an autonomous AI coding tool known as Devin, before expanding their AI suite by incorporating Anthropic's Claude models.
In the interview, Argenti remarked on the strengths of Claude in coding tasks, underlining the AI system’s capacity for logical reasoning and stepwise problem-solving. These capabilities extend beyond software development and prove advantageous in banking activities such as accounting and compliance, which require nuanced document interpretation and judgment calls.
While acknowledging the large workforce engaged in affected business areas, Argenti cautioned against premature speculation on staff reductions. Instead, he suggested that changes may primarily affect third-party service providers as AI adoption evolves.
This initiative reflects a broader trend in the banking sector towards leveraging AI to handle complex, labor-intensive processes. However, the ultimate impact on employment and operational structures within financial institutions remains to be fully realized as these technologies mature and their integration scales.