Goldman Sachs is exploring new ways to use artificial intelligence inside its operations. The bank is testing AI systems that can perform complex tasks on their own, reducing the need for large teams. This move aims to automate areas like accounting, compliance, and client onboarding, which are traditionally labor-intensive and rule-based.

Partnering with AI Startups to Develop Autonomous Agents

Goldman Sachs has been working with AI startup Anthropic for about six months. Engineers from Anthropic are working closely with Goldman’s teams to build AI agents powered by Anthropic’s Claude model. These agents are designed to handle detailed, multi-step processes that used to require human effort.

According to Goldman’s chief information officer, the AI systems have surprised staff with their capabilities. In early tests, the AI demonstrated the ability to reason through complex tasks such as accounting and compliance checks. This level of reasoning was unexpected and indicates a significant step forward for automation in banking operations.

Applying AI to Process-Heavy Back-Office Operations

Unlike many companies that use AI mainly for drafting texts or analyzing data trends, Goldman Sachs is focusing on automating back-office work. These areas involve a lot of rules, data review, and detailed analysis, making them difficult to automate fully in the past.

The AI agents are based on Anthropic’s Claude Opus 4.6 model, which can handle long documents and complex reasoning tasks. Early results suggest that these systems can greatly reduce the time needed for activities like client onboarding, trade reconciliation, and document review.

While the bank has not disclosed specific performance improvements, sources say that tasks which previously required significant human effort are now completed much faster. The goal is not to replace staff but to support them. Goldman Sachs sees these AI agents as tools to help employees manage high workloads and focus on more strategic work.

Overall, Goldman Sachs’s experiment with autonomous AI agents highlights a broader trend of financial institutions seeking practical AI applications beyond hype. This approach aims to improve efficiency and accuracy in critical, process-heavy functions.

Inspired by

• https://www.artificialintelligence-news.com/news/goldman-sachs-tests-autonomous-ai-agents-for-process-heavy-work/

Sources

• unsplash.com
• ai-expo.net
• techexevent.com
• techforge.pub
• techforge.pub