Leadership in the Age of GenAI: Mary Callahan Erdoes, Mike Urciuoli present Harvard case study
AWM OC members Mary Erdoes and Mike Urciuoli with Harvard Business School Karim R. Lakhani as well as AWM Technology and PB leaders Karen Donnelly, Clay Erwin, Robert Blanch, Vrinda Menon, Rod Thomas, David Ovitsky, Binita Narang and Caroline Sorensen for their Harvard case study presentation on AI's future in finance.
The hallowed halls of Harvard Business School have been the backdrop for many groundbreaking ideas. Last week, Asset & Wealth Management (AWM) CEO Mary Callahan Erdoes and AWM Chief Information Officer (CIO) and Chief Data & Analytics Officer (CDAO) Mike Urciuoli visited Harvard Business School (HBS) to discuss the benefits of advanced AI technologies in the financial sector.
They presented a newly-published HBS case study, “JPMorganChase: Leadership in the Age of GenAI,” developed in collaboration with Harvard’s Digital Data Design Institute. Authored by professors Iavor I. Bojinov and Karim R. Lakhani, along with Lead Case Researcher David Lane, the study explores JPMC's journey as an early AI adopter and its innovative use of generative artificial intelligence (GenAI) to transform operations, enhance efficiency and drive results. A first mover in AI.
The study, which is now part of HBS' core curriculum, details a March 2025 meeting where Erdoes along with JPMC CIO Lori Beer, JPMC CDAO Teresa Heitsenrether and Chief Risk Officer Ashley Bacon reflect on the impact and potential of GenAI for the 4,000 advisors serving high-net-worth Global Private Bank clients. After initially blocking access to ChatGPT to protect data security, JPMC has moved to steadily embrace the LLMs and the potential the technology offers.
Mary Callahan Erdoes was adamant from the beginning that AWM be a first mover in the space and spearheaded its adoption. She noted in the study, “There are lots of good reasons for the roadblocks, procedures and review committees. But, if we could do it faster, more creatively — that’s a whole new game.”
Heitsenrether agreed on its potential and benefits to the firm’s operations. “Every single person — from administrative assistants to code writers — would have it,” she said.
Since the release of ChatGPT in November 2022, the firm has implemented many large language models (LLMs), ensuring operational reliability, accuracy, security and compliance. The bank's GenAI toolkit, LLM Suite, is now deployed on the desktops of more than 200,000 employees, with more than half using it several times a day to help simplify and accelerate the completion of otherwise tedious tasks. These initiatives have already saved JPMorganChase nearly $1.5 billion through fraud prevention, service personalization, trading, operational efficiencies and credit decisions.
A transformative technology
Mary Callahan Erdoes and Mike Urciuoli take a photo with HBS students.
Having played a pivotal role in developing and implementing GenAI technologies at the firm, Urciuoli discussed how AWM immediately saw the extraordinary potential GenAI had for its business.
“Advice for our clients is rooted in language, and LLMs are a transformative technology in that domain," he said in the study. This sentiment was echoed during last week’s Harvard presentation, which underscored how the technology is already transforming our business.
The study also explored the development in house LLM applications, the firm’s commitment to data security, modernization and the promotion of tool adoption and development via gamification.
Coach Connect
The case study explored the Coach application, a GenAI tool embedded in the Connect platform for Private Bank advisors. Coach automates and scales the delivery of accurate, complete and easy-to-access information, empowering advisors to be able to deepen client relationships and expand their client rosters by 50%.
Emphasizing the importance of GenAI in enhancing advisor-client interactions, Karen Donnelly, former Private Bank Advisor and now Head of Digital for the U.S. Private Bank (USPB), said in the study: "If someone asks me to explain my role, I tell them that our team exists to make the advisor experience better. They want smart reasons to call clients. They want to cover more clients to have more revenue claim. GenAI lets us do more business, better business, more things faster, with more leverage."
In the study, USPB Head of AI and Innovation Mohamed Khalifa emphasized how the internal cost of expertise for basic information would be zero because advisors could get everything they need from Coach themselves, personalized for their clients' unique needs.
Data security and compliance
As JPMC is a market leader in a highly regulated and competitive industry, executives carefully considered how to best harness its capabilities before introducing it to the firm, ultimately deciding it would be best to build an internal LLM platform.
"Since data is our key differentiator, we had to ensure that our data was not being used to train external models,” said Heitsenrether. “We also had to design a system to comply with still-evolving regulatory requirements across jurisdictions."
AWM CEO Mary Callahan Erdoes presents to a class at Harvard.
IT and data infrastructure modernized
The study also noted how the firm’s strategic approach to AI adoption is further exemplified by a focus on creating a flexible, common tool chain for software development.
"We invest $17 billion in technology today – that was $9 billion when I joined J.P. Morgan in 2014,” said Beer. “We have integrated and substantially modernized our IT and data infrastructure."
Beer gained operating committee support by highlighting the potential commercial value external IT innovations could provide if applied to JPMC operations. “We laid the data on the table and showed the need for significant investment in an ongoing way, given the speed of evolution in technology today,” she said.
HBS students listen to Erdoes' presentation.
AI days and competitions: Rewarding inspiration
To boost engagement, Erdoes and Urciuoli organized AWM "AI Days.” During these events, cross-functional teams from the U.S., India and the U.K. presented demos of new tools and ideas, sparking inspiration and momentum across the business. The second AI Day featured 100 use cases, many with prototypes underway.
"Collaborative events like these inspire innovation and user adoption," said Urciuoli.
Following two successful AI Days, a global “Shark Tank”-style competition evaluated 500 GenAI proposals from early career analysts, with JPMC committed to developing the winning idea.
"This shows that early career ideas matter," said Donnelly of the success of the activity.
Employees were encouraged to attend training sessions showcasing how different teams and individuals were using the tools. Additionally, new GenAI innovations and innovators were spotlighted in employee communications. "Peer competition was an adoption mechanism," Urciuoli added. “We showed statistics on who was using the technology, how often, and by group.”
Erdoes credits this bottoms-up, iterative approach for AWM’s success in being an early AI adopter. “If you don’t make AI fun and personal, no one will adopt it at scale,” said Erdoes during the presentation at Harvard.
Unlock your wealth's full potential
We can help you navigate a complex financial landscape. Reach out today to learn how.
Contact us
