OpenAI has recruited more than 100 former investment bank analysts to participate in a project code-named Mercury. Its goal is to train artificial intelligence to build financial models and prepare presentation materials, writes Bloomberg expert Matt Levin.

The project involves specialists with experience at JPMorgan, Morgan Stanley and Goldman Sachs. They develop models in Excel and create tips for the AI, including adherence to industry formatting standards, such as italicizing percentages and proper margin formatting.

Participants get paid $150 an hour to work on models for a variety of transactions, including restructurings and initial public offerings, according to Bloomberg’s source. The company is giving experts early access to AI designed to replace entry-level tasks at banks.

Mercury’s project aims to automate routine functions for junior analysts. Once trained, the AI will be able to create models and presentations without the need for management review, which could reduce the need for junior staff and change the traditional career system in investment banking.

According to media reports, most junior analysts stay in banks for no more than two years. OpenAI is capitalizing on this trend by bringing in specialists who are already planning to leave the industry for accelerated AI training.

The project demonstrates the company’s commitment to bringing AI technology into investment banking to improve the efficiency of routine processes. Mercury remains an internal initiative of the company, and the results are not yet publicly disclosed.

Earlier, it became known that OpenAI reached a market valuation of $500 billion. This happened after the AI giant concluded a deal in which current and former employees sold shares to investors for about $6.6 billion.

Recall, we wrote that OpenAI introduced the ChatGPT Atlas browser.