JPMorganChase is gaining quantifiable efficiency through its AI-driven LLM Suite.  

The $3.6 billion bank rolled out the tool in the summer of 2024 and is seeing wide adoption in addition to the efficiency gains, Chief Data and Analytics Officer Teresa Heitsenrether said today at the Evident AI Symposium in New York. Evident AI is a data service provider and AI think tank based in New York City. 

Nearly 150,000 of JPMorgan’s 250,000 employees use the platform daily, she said. 

“Employees using the tool have reported an average of four hours of extra productivity each week,” Heitsenrether said.

Using LLM Suite, employees can create presentations, reference policies and access internal data sets, transcribe customer service calls and keep up with market research by providing internal and external data sets under one roof, she said. 

“It kind of ramped up gradually at first, and then quite a bit more steeply, and we’re seeing now a little bit more of a plateau,” Heitsenrether said. 

LLM Suite uses OpenAI and Anthropic base models, according to JPMorgan. 

The bank decided to create a multimodel platform because AI is changing quickly and the mode is intended to give employees “a choice of what is the right model for the use case that they are working on, whether it’s capability or cost effectiveness,” she said. 

Measuring ROI 

Measuring the return on investment of AI and generative AI is a tricky, Heitsenrether said. 

“We have traditional machine learning use cases that 100% create tangible value that we can measure that unequivocally continues to grow, on average, 30% or 40% a year,” Heitsenrether said.  

The bank has deployed traditional machine learning AI for fraud detection and anti-money laundering among other use cases and has seen efficiencies in those processes over the years, Heitsenrether said. 

According to Evident AI’s presentation today, JPMorgan says it found $2 billion of returns on its AI investments in fiscal 2024, primarily driven by traditional AI or machine learning. 

But Heitsenrether declined to be that specific today.  

“When you look at the generative AI aspect of it, and this roll out of LLM Suite, that’s very hard to measure,” she said. 

Heitsenrether said she expects that as the technology matures and is more widely used, quantifying its wins will become easier. 

Register here for early-bird pricing for the inaugural FinAi Banking Summit 2026, taking place March 2-3 in Denver. View the full event agenda here.