The Zhitong Finance App learned that OpenAI has launched a new version of ChatGPT to help Wall Street solve some labor-intensive tasks. This version of ChatGPT can be used to research companies, analyze financial data, and generate presentations that investment bankers need to survive. According to Nick Turley, VP of Product at OpenAI, this product, called ChatGPT for Financial Services, was released on Thursday. It is a customized version of its enterprise-grade product ChatGPT Work, which was jointly developed by OpenAI and “design partners” such as Morgan Stanley and Evercore. It uses OpenAI's newest and most advanced model, the GPT-6 Astra.
This announcement marks OpenAI's further expansion into an area traditionally dominated by entry-level bankers (so-called analysts and assistants) on Wall Street. Wall Street has been hiring these recent college graduates to research deals and produce promotional materials for decades. Furthermore, it also shows OpenAI's continued efforts for the upcoming and high-profile initial public offering (IPO) and its determination to continue to advance in the field of enterprise-grade products.
“We're actually teaching ChatGPT to do research like an analyst and provide arguments for its own conclusions like an analyst,” Turley said during a briefing to launch the new product.
Over the past year, OpenAI has been trying to win over customers in a highly competitive enterprise market, trying to resist competitors including Anthropic and Google (GOOGL.US). Last year, Anthropic unveiled its tailored solution for Wall Street — Claude for Financial Services.
OpenAI chief financial officer Sarah Friar told investors in August that the company's enterprise business revenue had surpassed the consumer business, which grew rapidly after the launch of ChatGPT in 2022.
Turley told reporters at the briefing that OpenAI plans to release customized solutions for “multiple industries” other than financial services.
In a live demo of the new product, Turley showed how the platform analyzes potential M&A targets, extracts financial data from industry-standard data sources, and creates formatted PowerPoint presentations based on the bank's pre-set formatting guidelines. Turley said, “It's easy to make beautiful slides, but it's much harder to make really meaningful slides. To reach the current level, ChatGPT must select relevant peer companies, import price data into a spreadsheet, check the chart against the data, and explain the reasons for the sell-off and rebound.”
Banking industry disrupted?
Compared to the product it's based on, ChatGPT Work, the biggest difference in this version is that it can obtain native data access from LSEG, Daloopa, and PitchBook to provide the system with information such as financial statements, profit reports, etc., and automatically access users' existing data subscriptions.
Other features tailored specifically for the financial industry include a citation function that allows users to trace data sources (such as original documents and audit charts), and administrative control over sensitive transaction materials.
Although Turley said that the latest version of ChatGPT, which was initially aimed at investment banking and equity research, is in high demand, he declined to reveal the names of banks that have signed up to use the software.
When asked if the latest version of ChatGPT would reduce investment banks' need for junior bankers, Turley described it as an efficiency boost aimed at maximizing the productivity of every employee.
“If you look into the lives of analysts or bankers, you'll find that, depending on the industry, they work up to 100 hours a week,” Turley said. “I think just as Microsoft Excel changed the entire industry, enabling them to produce better analytical reports faster, you'll see similar technology playing the same role.”
However, for an industry that has long been built on a strict apprenticeship model, this product has also raised some fundamental issues. If generative artificial intelligence can complete multi-step tasks such as research and formatting promotional materials within minutes, Wall Street will be forced to rethink how to train the next generation of deal makers, and how many such talents are needed.
Last month, Goldman Sachs partner Chris Churchman warned that automating tasks used to train junior bankers could cause a “cognitive decline” for the next generation of financial practitioners. Churchman is responsible for one of the bank's flagship artificial intelligence projects.
“Reasoning skills are still important,” Churchman said at the time. “You still need to reason about questions and structure them into arguments, and now we outsource the reasoning work.”