Sudipto Srivastava, Chief Data and Analytics Officer at Montefiore Health System

In this episode of the Becker's Healthcare Podcast, Laura Deirda interviews Sudipto Srivastava, Chief Data and Analytics Officer at Montefiore Health System, about the transformative role of AI and data in modern healthcare. Srivastava shares how Montefiore, a 13-hospital academic medical center serving over 7.5 million encounters annually, is leveraging AI across clinical, operational, and research domains—ranging from Ambient Voice AI scribes that reduce clinician documentation burden to AI governance frameworks evolving from Governance 1.0 to 2.0 and planning for 3.0. He emphasizes that AI is everyone’s responsibility, not a siloed function, and highlights key initiatives like auto-tagging cancer types with Gen AI, partnerships with Dandelion Health for de-identified data research, and deep integration with Epic, ServiceNow, and Workday. Srivastava also discusses critical challenges: balancing rapid innovation with risk mitigation, filtering signal from noise in AI hype, protecting patient data, and managing organizational change amid resistance. He stresses the importance of measuring AI’s real-world impact and retiring underperforming tools to avoid tech debt. Looking ahead, he identifies selecting the right problems to solve amid limited resources as the biggest challenge, while envisioning long-term organizational growth through redefining workflows and training models—such as using AI to broaden radiology residents’ interdisciplinary perspectives. The conversation underscores a data-driven, mission-aligned approach to AI that prioritizes sustainability, equity, and human-centered innovation in healthcare. Key takeaways include: (1) AI governance must evolve iteratively with input from all stakeholders; (2) Success requires measuring outcomes and retiring underperforming tools to avoid tech debt; (3) The most impactful AI initiatives align with real clinical and operational problems, not just technological novelty; (4) Redefining workflows—especially in training and interdisciplinary collaboration—offers a unique opportunity for long-term organizational growth; (5) Leadership must balance innovation with risk, especially around data privacy and vendor accountability; (6) AI adoption succeeds when driven by frontline staff and grounded in clear hypotheses; (7) Avoiding the 'addition without subtraction' trap is critical for sustainable tech integration; (8) Strategic prioritization of high-impact problems is essential in resource-constrained environments.