Mark Birkhead, firmwide chief data officer of JPMorganChase, said the bank is focused on a multi-year effort to pull humans out of the cycle for fixing data. In 2024 the bank set up a firmwide chief data and analytics office so that all its data initiatives were under one umbrella, led by Teresa Heitsenrether, chief data and analytics officer. She reports to chairman and chief executive, Jamie Dimon, and is part of the operating committee. Birkhead’s role as chief data officer is to oversee the team that sets the central data strategy for the firm in partnership with the business. Birkhead explained that the bank operates in almost 100 countries and across its consumer businesses, investment banking, asset and wealth management and payments. As a result, more than an exabyte of data moves across the firm in any day in many forms , including structured and unstructured data, voice and video files. An exabyte of data is so large that it has been estimated that all the words ever spoken or written by humans in every language since the beginning of mankind would fit on five exabytes. “Our data strategy centers on how we can best deliver all types of data assets and curate them in a way that is discoverable, highly accurate, and highly governed and controlled,” added Birkhead. “That last part is really critical for us as a bank because our customers count on us to keep their information private.” The group has built many data governance systems to ensure data risks are managed to respect privacy and is now investing in bringing them together into a central platform. A critical focus is modernizing the group’s data so that it can be published in a way that is consistent and understandable by large language models (LLMs) used in artificial intelligence and generative AI models. JPMorgan Chase has created an LLM Suite, which provides employees with access to LLM models. There are currently 200,000 users on LLM Suite, according to Birkhead. He continued that the bank is “particularly excited” about AI agents systems with more reasoning capabilities. Another focus is launching more data products and ensuring they are interchangeable, interoperable, and reusable. The final focus is ensuring that data is available in milliseconds.