Goldman Sachs CEO says bank will automate like a human assembly line

Goldman Sachs is signaling that the most routine parts of banking are no longer guaranteed apprenticeship work. When John Waldron called the firm a “human assembly line,” he was spelling out a future in which analysts and associates do less repetitive production work and more checking, judgment and client-facing problem solving.

Waldron, Goldman’s president and chief operating officer, made the remarks in a CNBC interview on May 12, 2026. He said the firm’s use of artificial intelligence is letting Goldman scale “without requiring much more hiring,” and argued that banks have not automated to the same extent as manufacturing. The message for younger bankers is blunt: the work that once justified long hours, from packet assembly to process cleanup, is becoming easier to digitize than to defend.

The first jobs most exposed are the ones closest to the factory floor of finance. Goldman has been working with Anthropic for about six months on AI agents for trade accounting and client onboarding, two areas heavy on standardized steps, documentation and exceptions. CNBC reported the bank expected to launch those agents soon. If those tools work as planned, the leverage shifts away from people who move data from one system to another and toward employees who can supervise outputs, catch errors, and handle the cases the software cannot resolve.

That does not mean a mass cull. Goldman’s broader plan, laid out by chief executive David Solomon in October 2025, was to reorganize the firm around generative AI while constraining headcount growth rather than slashing jobs in the near term. Reporting around Goldman’s One Goldman Sachs 3.0 memo said the bank expected limited role reductions and still anticipated a net increase in headcount by the end of 2025. Even so, the direction of travel is clear: the firm wants more output per employee, not more employees.

Goldman’s 2025 annual report put hard numbers behind the strategy. The bank reported net revenues of $58.3 billion, diluted earnings per share of $51.32 and return on equity of 15.0%. It also said it had more than 1.1 million experienced-hire applicants in 2025, while its workforce stood at 47,400 employees at year-end. Over six years, firmwide net revenues rose roughly 60% as historical principal investments fell by more than 90%, from about $64 billion to $6 billion, a sign that Goldman wants a more capital-light, more automated business.

For Goldman people, the career implication is that the premium will rise on skills manufacturing models reward too: process design, quality control, exception handling and the ability to manage systems at scale. Analysts who can validate AI output, associates who can redesign workflows and managers who can turn technology into cleaner execution will gain influence. The bank’s next productivity push is not just about replacing tasks. It is about deciding which bankers still own the line.