AI can draft 95% of a Goldman Sachs IPO prospectus in 10 minutes

About a decade ago during some of my keynotes about Artificial Intelligence (AI) and automation I said if you knew the process flow of a job or task then it was possible to automate all of it in time, and I used Goldman Sach’s 160 step IPO process as an example of a large very expensive process that would one day be fully automated. And now many many years later it’s happened – almost.

Goldman Sachs CEO David Solomon said investment banks could use AI to cut down on rote tasks like writing financial documents. Bankers used to have to spend weeks at a time drafting documents to be submitted for public filings. But at the Cisco AI Summit in Palo Alto on Wednesday, Solomon said that 95% of an S1 filing, which is the form a company files with the Securities and Exchange Commission when it goes public, can now be completed by AI in just a few minutes.

The efficiency gains are notable, considering it used to take a six-person team two weeks to complete the same task, Solomon said at the conference, according to the Financial Times.

Writing documents for the SEC is the exact sort of work that public and private companies of all sizes have historically relied on investment banks to advise them on. Investment banks often had the deep financial and legal expertise to properly navigate the intricacies of complex financial transactions that would have to be reported to regulators. With the rise of AI, investment banks like Goldman are rethinking their value proposition to clients.

“The last 5% now matters because the rest is now a commodity,” Solomon said.

Solomon’s latest remarks are in keeping with his past comments that part of Goldman’s focus with AI was its ability to help employees work more efficiently. AI-driven “Productivity [is] the ability to take smart people and give them tools so they can do more, do more quickly, help our clients think about things in different ways,” he told CNBC in May.

Over the past several years, Goldman Sachs has been changing the makeup of its workforce to become more focused on tech. As far back as 2018, a quarter of Goldman’s employees were computer engineers. Currently out of 46,000 employees, 11,000 are engineers, Solomon said. That is roughly 23% of Goldman’s employees, meaning that the share of engineers at the company has stayed roughly the same over that time.

In 2023, when AI was really proliferating across the corporate world after the launch of OpenAI’s GPT-3.5 in November 2022, a talent war broke out among Wall Street’s biggest banks like Goldman, Citigroup, and Wells Fargo. Each looked to poach the other’s best engineering talent to fill out their own ranks. Banks also created new executive-level roles to oversee AI across their entire organizations.

The hiring frenzy has meant that the pace of innovation on Wall Street has only accelerated. Goldman is not the only bank using AI in substantive ways across its operations. JPMorgan has a suite of internal AI chatbots for day-to-day tasks and AI models to help portfolio managers spot investment opportunities. Morgan Stanley has rolled out similar AI-powered tools for everyday productivity, including one called Debrief that specializes in taking notes during client meetings. Meanwhile, Wells Fargo has a customer service AI chatbot for its retail customers.