OpenAI has launched ChatGPT for Financial Services, a version of its workplace product designed to handle research, financial modelling and pitchbook preparation traditionally carried out by junior investment bankers and equity researchers.
The service uses the company’s latest model, GPT-6 Astra, and has been developed with Morgan Stanley and Evercore. It is initially focused on investment banking and equity research, where analysts often spend long hours gathering figures, comparing companies and preparing presentations for clients. ([openai.com](https://openai.com/index/introducing-chatgpt-financial-services/))
OpenAI vice-president of product Nick Turley said the aim was to teach the system to work “like an analyst”, including supporting its conclusions with evidence.
In a demonstration, the platform assessed a potential takeover target, selected comparable companies, collected market data, checked the figures and produced a PowerPoint presentation using a bank’s own templates and style guide.
The system includes financial information from providers including Daloopa, PitchBook and LSEG News, covering company fundamentals, financial statements and earnings transcripts. It can also connect users to existing subscriptions, including services from S&P Global, LSEG, MSCI, Dow Jones Factiva and Moody’s. ([openai.com](https://openai.com/index/introducing-chatgpt-financial-services/))
OpenAI said users would be able to trace figures and claims back to source material, while firms would have controls to manage access to sensitive deal documents, monitor activity and create barriers between teams handling confidential information.
The product is available to eligible financial institutions and allows administrators to publish Excel, Word and PowerPoint templates so that research notes, valuation models and pitchbooks can be produced in a firm’s established format. ([openai.com](https://openai.com/index/introducing-chatgpt-financial-services/))
ChatGPT for Financial Services raises questions over junior bankers
Mr Turley said the new service should be viewed as a way of increasing the productivity of existing staff rather than eliminating entry-level roles.
“If you study the life of an analyst or of a banker, depending on the industry, they’re working 100-hour weeks,” he said. He compared the technology with Microsoft Excel, which enabled bankers to produce more detailed analysis in less time.
OpenAI has not identified the banks that have signed up to use the product, although it said there had been strong demand. The company is also preparing tailored products for other industries as it seeks to expand its enterprise business.
The move comes as technology companies compete to establish themselves in financial services. Anthropic launched its own finance-focused offering last year, with tools for market research, financial analysis and the creation of documents and spreadsheets. ([anthropic.com](https://www.anthropic.com/news/claude-for-financial-services?utm_source=openai))
For Wall Street, the more difficult question is what happens to the apprenticeship system built around junior bankers carrying out repetitive but formative work. If software can complete research and presentation tasks in minutes, banks may have to reconsider both the size of entry-level teams and how new financiers learn their trade.
Chris Churchman, the Goldman Sachs partner overseeing one of the bank’s flagship artificial intelligence projects, warned last month that automating tasks used to train junior bankers could lead to “cognitive atrophy”.
“Reasoning is still important,” Mr Churchman said. “You still need to reason about [problems] and structure it into an argument, and now we’re delegating reasoning.”
