Goldman's engineers have a new challenge: turning AI agents into firm insiders
Marco Argenti, the firm's chief information officer, said engineers are now focused on transferring "tribal knowledge" to AI agents.
Chief Information Officer Marco Argenti of Goldman Sachs has a new challenge on his hands - transforming artificial intelligence (AI) agents into insiders familiar with the bank's culture and standards. As more than 12,000 Goldman developers now utilize AI tools, Argenti and his team are focusing on creating AI tools that are specifically tailored to Goldman's unique requirements.
To accomplish this, the bank's developers must teach AI agents its "tribal knowledge," the subtle nuances that define Goldman's engineering culture. This involves mentoring the AI in aspects such as the company's data standards, security protocols, and design principles. AI agents that possess this knowledge can produce better work more quickly for engineers to review.
Goldman Sachs has allocated around $6 billion to AI last year and faces increasing pressure to demonstrate the productivity of these investments. CEO Jamie Dimon has stated that spending on AI is essential to remain competitive. The bank's engineers are now tasked with teaching AI agents the "tricks and tribal knowledge" that the technology doesn't inherently understand.
Goldman has developed skills - reusable bundles of instructions - to capture developers' technical knowledge about the bank. These skills include migration procedures for cloud deployment and specific cloud migration strategies applicable to Goldman. The skills allow AI tools to be more helpful to those responsible for cloud migration, similar to how an experienced employee would be more helpful than a new hire.
While AI tools capture the "written" rules, the "unwritten" ones are more challenging to systematize. As part of the effort, Goldman has undertaken internal process evaluations, including interviewing employees and analyzing outputs. The bank continually updates its skills as its internal processes evolve, creating a cycle of improvement in which the agents' code becomes better and more specific to Goldman.
Argenti also mentioned that mentorship at Goldman is changing. While junior employees typically seek guidance from managers, they are now mentoring firm veterans on AI use in both structured and unstructured ways. This shift is expected to have broader implications for other lines of business within the bank.
Written by urgent.news from Business Insider's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
