Dubbed LLM Suite, the bot is designed to help with writing, idea generation and summarising documents, executives told staff. “Think of LLM Suite as a research analyst that can offer information, solutions and advice on a topic,” an internal memo seen by the Financial Times says. Already at the desks of 50,000 staff, the next phase of the roll out has placed the product in the hands of employees of the lender’s asset and wealth management division. Speaking at the bank’s investor day in May, Mary Erdoes, head of the asset and wealth management division, said: “This year, everyone coming in here will have prompt engineering training to get them ready for the AI of the future.” Company president Daniel Pinto said at the event that the technology will have a “very, very” big impact on the bank’s 60,000 developers and 80,000 operations and call-center employees. In his April letter to shareholders, JPMorgan chief Jamie Dimon likened the arrival of AI to the development of the printing press and the steam engine. "In the future, we envision GenAI helping us reimagine entire business workflows," wrote Dimon, adding that the technology "has the potential to augment virtually every job".

In May, the bank announced that it was using OpenAI’s GPT-4 model for a range of thematic investment baskets, identifying keywords associated with a theme before putting them into a natural language processing model that scans news articles to identify companies involved in the area. The Wall Street giant now has more than 2000 AI and machine learning experts and data scientists, helping it to bring in over 400 use cases for the technology in areas such as marketing, fraud and risk. A recent study by Evident found that JPMorgan has steadily increased its share of AI research output from 30% in 2018 to 45% in 2023 amidst heavy investment in AI talent. The bank employs over 200 AI researchers, more than four times that of its nearest rival Royal Bank of Canada and nearly half of the talent pool available across the Top 10 banks for AI research.