12,000 developers and thousands of operations affected! Goldman Sachs teams up with Anthropic to achieve full automation of accounting and compliance operations

Goldman Sachs is taking a significant step toward embracing artificial intelligence.

According to the latest news, they are deploying Anthropic’s Opus 4.6 model across the board, aiming to automate accounting and compliance roles. This marks a major advancement in the financial industry’s use of generative AI to handle core back-office operations.

Six Months of Deep Collaboration with On-Site Engineers

This transformation did not happen overnight. It is reported that Anthropic’s engineers have been stationed inside Goldman Sachs for up to six months. Through this embedded, in-depth collaboration, both parties jointly developed a system referred to by Goldman Sachs as “digital co-workers.” The primary mission of these AI systems is to assist human employees in handling large volumes of repetitive, process-heavy tasks.

Claude Model Demonstrates Remarkable Reasoning Ability

The core of this new system is based on an LLM (Large Language Model) agent. It has powerful document processing capabilities, able to read bundles of transaction records and complex policy texts. More importantly, this AI can strictly follow step-by-step rules, autonomously determining the next action—whether to process directly, flag anomalies, or route to relevant personnel for approval.

Goldman Sachs revealed that this collaboration brought an unexpected surprise: Claude’s capabilities extend far beyond coding. The bank found that the model’s logical reasoning style in programming is also perfectly suited for rule-based accounting and compliance work. Even in complex scenarios involving large amounts of text, data tables, and various exceptions, Claude can still perform logical judgments and processing with ease.

Efficiency Gains and Workforce Strategy Adjustments

Goldman Sachs has high expectations for the deployment of this AI system. It is expected to significantly shorten client review cycles and reduce long-standing issues in transaction reconciliation, thereby greatly improving operational efficiency.

Regarding the highly discussed employment impact, Goldman Sachs currently maintains a cautious stance. The bank states that the immediate result of introducing AI automation will be a slowdown in future hiring growth, rather than large-scale layoffs.

According to reports, this change affects over 12,000 developers and thousands of operations staff, who rely on Claude’s advanced reasoning capabilities to manage a portion of the bank’s $2.5 trillion in assets.

Unlike standard chatbots, these autonomous agents utilize Claude Opus 4.6’s massive 1 million token context window to process complex financial data in real time.

Goldman Sachs states that the AI coding assistant has increased developers’ productivity by over 20%, reducing thousands of hours of manual work each week.

By using Claude, the bank has shortened onboarding times for new institutional clients by 30%. These agents match global databases with internal compliance rules to complete “Know Your Customer” (KYC) and anti-money laundering checks. This extensive data workflow ensures every transaction complies with strict regulations.

Goldman Sachs plans to expand its application to employee behavior monitoring and investment banking material preparation.

In Conclusion

When a highly regulated company like Goldman Sachs applies agent-based AI to daily operations, it powerfully demonstrates a real need beyond simple chatbots.

This move may prompt companies to invest more in model providers, cloud platforms, and consulting firms that help integrate and manage these systems.

At the same time, it puts pressure on outsourcing companies and business process firms that profit from large-scale document review and verification.

Banks rarely adopt off-the-shelf automation systems directly; instead, they perform extensive customization. Controls, audit trails, and clear accountability are essential.

Placing engineers directly within Goldman Sachs teams indicates that the real advantage lies in integrating AI into traditional systems and compliance frameworks.

If this approach proves effective, more “co-constructing” collaborations could change how enterprises sell, manage, and scale AI solutions.

Source: AI Cambrian

Risk Warning and Disclaimer

Market risks are present; invest cautiously. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situations, or needs. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. Invest at your own risk.