Hard Lessons

Suman’s work sits at the collision point of AI ambition and enterprise reality. From Aetna and Highmark to Merck and Pfizer, he has repeatedly been asked to build new data science and AI capabilities inside complex healthcare organizations, where fragmented data, regulatory pressure, and internal politics shape what can actually scale. In this conversation, he traces how the field evolved from early machine learning and open-source analytics into today’s generative AI wave, and why the hardest part has remained surprisingly consistent: getting people, incentives, leadership, and adoption aligned. His lessons move beyond the usual AI hype, offering a grounded playbook for building durable teams, earning executive cover, designing products people want to use, and turning AI from a spotlight initiative into real organizational muscle.

Across Nike and American Express, Elaine Barsoom had a front-row seat to one of the hardest problems in corporate innovation: how to turn promising technology into something a large organization can actually absorb. She reflects on the moments that defined that work, from helping Nike pursue a direct-to-consumer shift through external partnerships and startup pilots, to standing up an AI center of excellence as generative AI began reshaping the enterprise agenda, to building and scaling a joint venture at Amex that reached real commercial traction. Along the way, she offers a clear-eyed view of why so many innovation efforts stall, why pilot success rarely tells the whole story, and why executive sponsorship only matters if it holds when adoption gets difficult. The result is a practical lesson in organizational readiness, AI governance, and the human work required to make transformation stick.