- Goldman Sachs CEO David Solomon spoke at the Cisco AI Summit this week.
- He said AI was changing processes like drafting IPO filings and analyst research.
- Efficiency gains rely on the execution of employees changing their processes.
One of Wall Street’s top bosses just gave a revealing look into how AI is changing the lives of bankers and analysts in its investment bank.
Ten years ago, when Goldman wanted to win the business of a company going public, it would appoint a team of about half a dozen people who over two weeks would draft a prospectus known as an S-1, a significant regulatory document that details the business, financials, and risk factors, among other things. The idea was to show would-be clients all the thought and legwork Goldman bankers had already put in.
“Now you can basically have something that’s 95% of the way there in a few minutes,” CEO David Solomon said at Cisco’s AI Summit on Wednesday. If the 95% is “now a commodity,” he said, the remaining 5% matters a lot because it’s the margin where Goldman can get an edge.
It’s a big deal, as Goldman is one of the top banks that take companies public, along with Morgan Stanley and JPMorgan. It’s also just one of the ways Goldman is using AI to reduce grunt work and move more efficiently.
Solomon described a road map for how Goldman Sachs was strategizing around AI. He said that getting engineers to free up capacity by being 30% more productive with coding tasks was “first and most obvious,” given that Goldman has 11,000 engineers. Another priority is making better use of Goldman’s data, which includes tracking the firm’s trades over the past 40 years and making it available to clients.
The third, and perhaps most visible and directly client-facing, is deploying AI in the investment-banking business. Enabling the bank to do more work by giving workers a kind of information superintelligence would boost the already booming firm, which brought in more than $53 billion in 2024.
Beyond using AI to draft IPO prospectuses to court potential clients, Solomon said Goldman was “focused on how we can completely change kind of what you’d call the material preparation” involved in investment banking. That includes preparing bankers for client meetings and getting information to clients to better make investment decisions. He said Goldman was also building an “investment-banking copilot” with the bank’s own data. Copilots typically describe AI tools that help workers be more productive by drafting text, analyzing information, and suggesting ideas.
It’s not just bankers who can anticipate changes to their daily processes. The Goldman CEO also talked about the potential for AI to shake up analyst workflows in equity research.
He said one of the most significant things analysts and their dedicated teams do is report on companies and feed that information into models, often to gauge a company’s trajectory, growth, and risks.
“Obviously that can all be automated now with this technology,” Solomon said.
But that doesn’t mean those jobs will be replaced with technology. “You really need the analysts, and you need smaller teams, and you need a horizontal engine that basically does all that work for everyone as opposed to individual pods for every single industry,” Solomon said.
The success of these AI deployments will rely in part on the execution of change management.
Solomon said that generating efficiencies and cutting down on redundancies in the equity-research example would be “a massive process change.”
“It’s hard, because people don’t want to change their process,” Solomon said. “They like their team. They like the fact they have complete control.”
Read the full article here
