According to Bloomberg, the American AI giant OpenAI has assembled a team of over a hundred former investment bankers from major institutions such as JPMorgan Chase, Morgan Stanley, and Goldman Sachs for a confidential project called Mercury.

OpenAI’s ambition is to automate the tedious tasks traditionally assigned to junior bankers and young analysts. Project Mercury aims to equip AI systems with the ability to perform financial modeling, a critical step in investment banking work.

These AI tools could soon handle report generation, presentation preparation, and valuation modeling—time-consuming tasks often referred to as “grunt work,” the most repetitive part of the job.

Large-Scale Testing and Early Access

To achieve this goal, OpenAI has given project participants early access to its developing technologies to test reliability on real cases such as IPOs, restructurings, and mergers and acquisitions.

The former bankers involved write detailed instructions and design template models to train future AI tools capable of replicating—and even optimizing—the decision-making and calculation processes used in major financial institutions.

A Strategic Shift for OpenAI

This project reflects OpenAI’s desire to make its AI directly usable by businesses, particularly in analytically intensive sectors like finance, consulting, and law. After gaining public recognition with ChatGPT, the company now aims to demonstrate the operational value of its models in professional environments where accuracy, speed, and regulatory compliance are essential.

Balancing Innovation and Job Concerns

While this development promises a significant reduction in manual tasks, it also raises concerns about the future of employment in finance. Young analysts, often exhausted by demanding workloads, have mixed reactions: some welcome the relief, others fear job displacement. Many worry that technology could eventually replace part of their roles, fundamentally reshaping the structure of investment banking.