(US & Canada) | Creating an AI-friendly Culture Is a Key Trend for Data Leaders — JPMorganChase VP CCB

Meera Dugar, VP, Consumer and Community Banking (CCB) at JPMorganChase (JPMC), speaks with Alexandra Sawyer, Director at Guidehouse, in a video interview about challenges around creating a data and AI culture, balancing regulatory risk with AI adoption, and top trends data leaders should be looking out for.

Data leaders, according to Dugar, face unique challenges, primarily in fostering an environment and culture of AI and data awareness within their organizations. This requires a top-down approach to disseminate knowledge and real use cases for AI models. By doing so, they can encourage the adoption of AI technologies and offer customers initial experiences with these models.

Dugar emphasizes that collecting feedback from these initial experiences is crucial. It allows organizations to refine the models and create a robust human feedback loop. This process not only strengthens the knowledge base but also helps in minimizing false positives and ensuring data accuracy.

When asked about managing the regulatory risk around the application and adoption of AI models, Dugar stresses the importance of ensuring data security by making sure data is securely containerized and preventing any data leakage. She points out that when providing investment recommendations, it is crucial to maintain transparency and trust with customers. For instance, if a customer is advised to invest in a particular stock with a high return rate, it is essential to ensure they are not misled into fraudulent schemes like pump and dump strategies. This transparency, according to Dugar, is central to building strong customer partnerships and trust in AI-driven systems.

Further, Dugar emphasizes the need for being highly cognizant of customer needs and providing value through AI. She discusses the importance of starting with basic models and gradually progressing to the more sophisticated ones. This involves refining workflows with better filters, prompts, and responses, ultimately creating smarter models that offer unprecedented value.

By doing so, organizations can achieve significant breakthroughs, or "aha moments," enhancing both customer satisfaction and the perceived value of AI. According to Dugar, this approach not only fulfills regulatory requirements but also builds stronger, more trustworthy relationships with customers.

Speaking of top trends data leaders should be looking out for, Dugar emphasizea the importance of fostering an AI-friendly culture. This involves engaging subject matter experts and utilizing resources like newsletters and journals to stay informed. She points out the value of sources such as Michael Proksch's book "The Secrets of AI Value Creation" and publications like IBM's Think and Dell's IQ magazine. These resources can help individuals learn from the community and develop insights relevant to their specific domains.

Additionally, Dugar highlights the necessity of understanding data to identify accurate use cases that yield the best returns and provide significant value. She stresses the importance of maintaining accurate schemas and using clustering techniques to ensure response accuracy. Dugar also mentions the critical role of memory in large language models (LLMs) and the need for proper GPU infrastructure to support their memory-intensive requirements.

CDO Magazine appreciates Meera Dugar for sharing her insights with our global community.