Half of 46,000 Goldman Sachs employees now have access to artificial intelligence. By the end of this year, Mark Argenti’s chief information director expects even more that he will be able to use the II in the hope of improving their productivity – but not everyone is still in the firm.
“We have the whole organization that needs to be re-established and re-develop AI,” Argenti says. “But, I think, we were very, very deliberately in relation to managing people by managing people.”
The measured approach in Goldman Sachs, which ranks #35 on Fortune 500Displays Argenti’s opinion that AI technology develops and is still coming with a lot of uncertainty. The Goldman’s AI Group Group and the Risk and Control Group are working to determine which of dozens of AI’s proposals and how it can be done responsibly.
One examples is an experiment that Goldman does with Agentic AI, which is still not completely deployed. II agents are designed to work autonomously or with little human supervision, performing multi -stage reasoning or tasks. These agents can, theoretically, conduct compliance or help process customer operations. But AI agents also require some preparation and can hallucin, which leads to errors in the results they give. Goldman says he still evaluates what additional control he needs to be effective and safe to use Agentic AI.
Because Goldman works in a very adjustable sector, the industry historically prefers to create technologies, giving these institutions more control to protect sensitive financial data customers. This has changed with a cloud lift, and most recently generative and The vast majority Financial institutions have deployed at least one AI generative product, often cooperating with external suppliers.
Approximately one of the four Goldman Sachs employees is an engineer, and this group was the first, focused on Argenti when deploying AI generative instruments. Argenti gave these workers access to AI Coding Assistant tools, including GitHub Copilot and Assist Code Gemini. Goldman held competitions inspired by show -competition realistic television Shark For developers to share their most creative use of II.
Argenti measures the profitability of investment from these Copilot tools in several ways, including the frequency of use and speed of code, generated by GitHub and similar tools.
The widespread use of the Generative II in the company came with the launch of the GS AI assistant, which unfolded last year and expanded up to 10,000 employees, including bankers, merchants and asset executives. This Goldman tool will be available to almost all employees by the end of 2025, can summarize documents, design emails, analyze data and create personalized content.
The GS AI assistant was built as a multi -mode, using large linguistic models with Gemini, Openai and Llama, and Argenti studies LLM from other AI Hyperscalers. Argenti says he does not want to rely on only one supplier and gives the firm flexibility to use a model that may be better for coding, while the competitive offer is stronger in reasoning. Goldman is also factors in how easy LLM is to change and how expensive they are.
“All these opinions have given us to the fact that we want to continue to connect to these models,” Argenti says.
For workers who are not in engineering, Goldman monitors use rates and sends surveys to get feedback to further improve the GS AI assistant. The company sought to promote the champions from asset management and wealth, private banking business and trade-non-technologists to get a purchase. “People can be afraid or skeptical when you first manage technology,” Argenti says.
Argenti joined the firm as a partner and joint CIO in 2019 and fully assumed the role in 2022 after his co -author George Lee became co -chairman of the geopolitical and technological understanding of Goldman Sachs Global Institute. Prior to the arrival in Goldman, Argenti was vice -president on technology in Amazon Web Services for six years, and also played a leading role in a telecommunications company Nokia.
Many of his previous works in Goldman focused on the fact that the firm staff work at home as a result of a global pandemic. But he also wanted to move the technology culture to be less like the bank thinking about technology, which usually contributed to the creation of custom decisions for each individual division, and more like a technological giant that creates one tool that can be divided into the company.
This new way of thinking is displayed in the legend of data management platform and open source management, which is publicly launched in 2020. Goldman has built a legend internally 10 years before public launch, giving both technologists and non -technical users the ability to develop applications focused on data and get understanding from this data.
The data is the key component of the AInti strategy, which it calls a three-year chair, which also must represent the AI technology itself and the people who use it. Good quality data is necessary for the correct exit from LLM, but the change of behavior of people is equally important.
“It is about the possibilities of strengthening the best people, I think you will get the best results,” Argenti says.
John Kel
Send your thoughts or suggestions to CIO Intelligence there.
Originally this story was presented on Fortune.com
https://fortune.com/img-assets/wp-content/uploads/2025/03/M.Argenti-Headshot.jpg?resize=1200,600
2025-03-19 17:13:00
John Kell
