OpenAI’s Secret Project Mercury Enlists Wall Street Veterans to Automate Banking Tasks

OpenAI’s Covert Banking AI Initiative

OpenAI is reportedly developing artificial intelligence specifically designed to handle the tedious work performed by junior investment bankers, according to documents reviewed by Bloomberg. The secretive project, code-named Mercury, has enlisted former employees from major financial institutions including JPMorgan Chase, Morgan Stanley, and Goldman Sachs to help train its systems.

OpenAI’s Covert Banking AI Initiative
High-Priced Financial Expertise
The Banking Grind Culture
AI-Driven Recruitment Process
Wall Street’s Elite Participation
OpenAI’s Business Expansion Strategy

High-Priced Financial Expertise

Participants in Project Mercury are reportedly compensated at $150 per hour for writing prompts and constructing financial models covering various transaction types, sources indicate. The contractors are working on models for restructurings, initial public offerings, and other complex financial transactions, according to people familiar with the effort.

These financial experts have received early access to the AI system OpenAI is building, which aims specifically to replace entry-level tasks at investment banks, the report states. This approach underscores OpenAI’s strategy to develop industry-specific AI solutions by collaborating directly with domain experts.

The Banking Grind Culture

Investment banking analysts typically work upwards of 80 hours per week when handling live deals, primarily building detailed models in Microsoft Excel for mergers, acquisitions, and leveraged buyouts. The culture famously includes constant requests from senior bankers to make endless adjustments to PowerPoint presentations – a phenomenon that has spawned Wall Street’s “pls fix” meme.

While junior bankers have long complained about the drudgery, the rise of AI is now generating concerns about job security, analysts suggest. Multiple startups are already developing AI tools to assist with these banking tasks, but OpenAI’s direct engagement with Wall Street veterans represents a significant escalation in this trend.

AI-Driven Recruitment Process

The application process for Project Mercury involves almost no human interaction, according to sources familiar with the matter. Candidates first undergo a 20-minute interview with an AI chatbot that asks questions based on their resume, followed by testing on financial statement knowledge, and finally a modeling assessment.

The contractor role is reportedly flexible, requiring approximately one model submission per week. Participants receive instructions to write prompts in simple terms before executing the models, and they obtain feedback from reviewers before their work is integrated into OpenAI’s systems, sources indicate., according to technological advances

Wall Street’s Elite Participation

Project Mercury has attracted participants from prestigious financial institutions including Brookfield Corp., Mubadala Investment Co., Evercore Inc., and KKR & Co., according to the documents. The project also includes current MBA candidates from Harvard University and Massachusetts Institute of Technology, demonstrating OpenAI’s focus on recruiting top financial talent.

Participants are required to create models in Excel while adhering to industry formatting standards, including specific margin sizes and italicized percentages, maintaining the professional presentation standards expected in investment banking.

OpenAI’s Business Expansion Strategy

The project highlights OpenAI’s urgent push to make its AI technology more practical for businesses across consulting, finance, legal, and technology sectors, analysts suggest. Despite recently reaching a $500 billion valuation, the company has yet to achieve profitability, increasing pressure to develop commercially viable applications.

An OpenAI spokesperson stated that the company collaborates with various experts “to improve and evaluate the capability of our models across different domains,” noting that these specialists are “recruited, managed and compensated by third party suppliers.”

This development comes as the financial industry increasingly explores artificial intelligence applications to streamline operations, particularly in investment banking where junior staff traditionally bear the burden of repetitive analytical work.