How Goldman Sachs’ CIO is Preparing for AI Agents

Few investment banks have been able to stay at the forefront of the industry in the same way as Goldman Sachs. Now, the Wall Street icon is taking an early lead in the AI race, one that promises to reshape the industry and potentially challenge the dominant status of many leading financial institutions.

At least 10,000 bankers, traders and asset managers already have access to a GS AI assistant, a Generative AI assistant.Today, the system helps with important, but repetitive tasks, like summarizing emails or translating code. But as the capabilities of large language models (LLMs) continue to improve rapidly, the goal is for the technology to become an integral part of operations over the coming years.

However, alongside improvements in foundational models, Goldman Sachs Chief Information Officer Marco Argenti also has to navigate several key challenges to make the vision a reality, including getting GS AI to adapt to the institution’s culture.

“The AI assistant becomes really like talking to another GS employee,” Argenti told CNBC.“For the AI to have a very specific identity that reflects the tenets, the values, the knowledge and the way of thinking of the firm is extremely important.”

Below are key pieces of advice Argenti has shared throughout Goldman Sachs’ AI agent journey.  

On people:

• Transformation requires people: “The biggest friction points to adoption of AI in the enterprise is people, behaviours, the fact that you have muscle groups that you need to retrain.” – Argenti said on the “Goldman Sachs Exchanges” podcast.
• ‍A focus on human empowerment: “People are going to make a difference, because people are going to be the ones that actually evolve the AI, educate the AI, empower the AI, and then take action,” Argenti told CNBC.
• Find your disruptors: “You need to find people that have already crossed the bridge of accepting that a disruption is going take place and be very open to it.” Goldman was actively seeking out ‘disruptors’ as part of its AI rollout,” he said on the podcast

On process:

• ‍Orchestration is the goal: “As we progress, the second step is when you’re starting to have this agentic behavior, that is, ‘I’m completing a task on behalf of a Goldman employee, and I need to take a set of steps,’” Argenti told CNBC. “That’s where the model is going to start to do things like a Goldman employee, not only say things like a Goldman employee.”
• The new hybrid team: "The capabilities of AI models to plan and execute complex, long-running tasks on humans' behalf will begin to mature...this will create the conditions for companies to eventually 'employ' and train AI workers to be part of hybrid teams of humans and AIs working together,” Argenti wrote in a company blog.
• Embrace the digital HR: “Companies will reskill human managers to oversee a hybrid workforce. The role of human resources will evolve into a department for human and machine resources,” Argenti wrote in the blog.

On technology:

• Domain-specificity isn’t easy: “One of the challenges is to somehow inject the cultural traits, the leadership principles, the tenets, of the organization into the AI agents the same way as you do with humans,” Argenti told The Stack.
• ‍Your data is key: “Agents that you implement from the outside are going to be like employees that we hire the first day. And so the challenge is agents will get smarter and smarter, but not culturally smarter if you don't do something on that front,” he said on the podcast.
• ‍The application layer is the new battlefield: “Startups that are now ‘model-centric’ will shift towards building solutions that are model-agnostic, focusing instead on other aspects of AI such as compliance, safety, data integration, orchestration, automation, and user experience,” he wrote in the blog.