JPMorgan Chase Unveils Generative AI Assistant for Employees

JPMorgan Chase has recently introduced a generative artificial intelligence assistant to tens of thousands of its employees, marking the initial phase of a broader strategy to integrate this technology across the expansive financial institution.

The program, known as LLM Suite, is already accessible to over 60,000 employees, assisting them with tasks such as writing emails and reports. According to sources familiar with the plans, the software is anticipated to become as widespread within the bank as the videoconferencing tool Zoom.

Instead of developing its own AI models, JPMorgan designed LLM Suite to serve as a portal, enabling users to access external large language models — the sophisticated programs that power generative AI tools. The suite was launched using OpenAI’s LLM, the creators of ChatGPT, according to these sources.

“Ultimately, we’d like to be able to move pretty fluidly across models depending on the use cases,” said Teresa Heitsenrether, JPMorgan’s chief data and analytics officer, in an interview. “The plan is not to be beholden to any one model provider.”

ChatGPT Ban

More than a year after restricting employees from using ChatGPT, JPMorgan Chase has introduced its own version of OpenAI’s ChatGPT, called LLM Suite, tailored to the bank’s requirements. This move ensures that the bank’s proprietary data remains secure and is not exposed to external providers, according to Heitsenrether.

“Since our data is a key differentiator, we don’t want it being used to train the model,” she said. “We’ve implemented it in a way that we can leverage the model while still keeping our data protected.”

The LLM Suite has been broadly deployed across the company, with teams in JPMorgan’s consumer division, investment bank, and asset and wealth management business using it. The tool assists employees with writing, summarizing lengthy documents, solving problems using Excel, and generating ideas.

However, Heitsenrether, who was promoted in 2023 to lead the bank’s adoption of this cutting-edge technology, emphasized that simply providing access to the models is just the first step.

“You have to teach people how to do prompt engineering that is relevant for their domain to show them what it can actually do,” Heitsenrether said. “The more people get deep into it and unlock what it’s good at and what it’s not, the more we’re starting to see the ideas really flourishing.”

The bank’s engineers can also use the LLM Suite to integrate functions from external AI models directly into their applications, she added.

“Exponentially Bigger”

JPMorgan has been working on traditional AI and machine learning for over a decade, but the advent of ChatGPT required a strategic pivot.

Traditional, or narrow, AI performs specific tasks involving pattern recognition, like making predictions based on historical data. Generative AI, however, is more advanced and trains models on vast datasets to create patterns, resulting in human-like text or realistic images.

The number of applications for generative AI is “exponentially bigger” than previous technologies due to the flexibility of large language models (LLMs), Heitsenrether said.

The bank is testing many use cases for both forms of AI and has already put several into production.

JPMorgan is using generative AI to create marketing content for social media channels, plan itineraries for clients of the travel agency it acquired in 2022, and summarize meetings for financial advisors, she said.

The consumer bank uses AI to determine optimal locations for new branches and ATMs by analyzing satellite images and in call centers to help service personnel quickly find answers, Heitsenrether said.

In the firm’s global payments business, which moves more than $8 trillion around the world daily, AI helps prevent hundreds of millions of dollars in fraud, she noted.

However, the bank is more cautious with generative AI that directly interacts with individual customers due to the risk of chatbots providing incorrect information, Heitsenrether said.

Ultimately, the generative AI field may consolidate into “five or six big foundational models” that dominate the market, she predicted.

JPMorgan is testing LLMs from U.S. tech giants and open-source models to integrate into its platform next, according to sources familiar with the bank’s AI strategy.

Friend or Foe?

Heitsenrether outlined three stages for the evolution of generative AI at JPMorgan.

The first stage is simply making the models available to employees; the second involves integrating proprietary JPMorgan data to enhance employee productivity, which is the current stage at the company.

The third stage represents a significant leap, where generative AI becomes powerful enough to operate as autonomous agents performing complex multistep tasks. This would transform rank-and-file employees into managers with AI assistants at their command.

The technology is expected to empower some workers while displacing others, potentially altering the industry’s composition in unpredictable ways.

Banking jobs are the most prone to automation compared to other industries like technology, healthcare, and retail, according to consulting firm Accenture. AI could boost the sector’s profits by $170 billion within four years, according to Citigroup analysts.

People should view generative AI “like an assistant that takes away the more mundane tasks we would all prefer not to do, providing answers without the need to grind through spreadsheets,” Heitsenrether said.

“You can focus on the higher-value work,” she added.