From Supply Chain Analytics to Agentic AI: How AI is Transforming Supply Chain Decision-Making

a futuristic data dashboard featuring bar and line graphs
Publication Date
September 15, 2026
With AI tools able to predict, adapt, and make autonomous decisions before a disruption even occurs, organizations struggling to implement AI technology into their supply chain decision-making strategy risk getting left behind.

The MITx MicroMasters Program in Supply Chain Management held its latest webinar, From Supply Chain Analytics to Agentic AI: How AI is Transforming Supply Chain Decision-Making on Thursday, September 10, 2026.

The online session, hosted by Dr. Eva Ponce, Director of Online Education at the MIT Center for Transportation and Logistics, featured guest speaker Vijay Sankararaman, Chief AI Officer at Novant Health and former technology leader at Lowe's and Walmart. The discussion explored the rapid evolution of supply chain analytics—moving away from reactive, descriptive dashboards focused purely on cost-cutting, toward advanced machine learning models capable of processing massive, complex variables to drive proactive decisions.

Dr. Eva Ponce of MIT CTL and Vijay Sankararaman of Novant Health discuss the evolution of supply chain analytics from  predictive analytic and machine learning to generative and agentic AI in this engaging session.

Drawing on his cross-industry experience, Sankararaman outlined the technological progression from predictive insights to autonomous execution. He shared how Lowe's utilized ensemble machine learning models to navigate constrained appliance supply chains by predicting highly nuanced consumer demand to optimize last-mile shipping. Now in the healthcare sector at Novant Health, Sankararaman utilizes Agentic AI to streamline complex procurement processes. By deploying AI agents to analyze thousands of overlapping medical contracts, his team can identify pricing variances, recommend standardizations, and draft supplier communications, all while ensuring human clinicians remain firmly in the loop.

Key Takeaways

  • The AI Capability Spectrum: Sankararaman defined the evolution of AI across three distinct stages: 
    • Predictive AI outlines what will happen. 
    • Generative AI creates net-new content, like drafting supplier negotiations.
    • Agentic AI bridges the gap by taking autonomous action based on that context.
  • When to Delegate to Agents: Not all tasks should be automated. Agentic AI is best suited for decisions that are routine, reversible, and have bounded costs. High-stakes, strategic, or irreversible decisions still strictly require human judgment and oversight.
  • The "Quantitative Grind" is Shrinking: As AI takes over routine data processing and batch runs, the manual day-to-day grind for professionals will decrease. This represents a shift in work rather than an elimination of jobs, elevating the value of human judgment.
  • The Foundations Still Matter: As technology evolves, a strong foundation in analytics, forecasting, optimization, and supply chain fundamentals remains essential. AI builds on these foundations, it doesn’t replace the need to understand the underlying business and supply chain problems. Both Sankararaman and Ponce emphasized the importance of critical thinking and business acumen to ask the right questions, evaluate AI-generated insights, and ensure technology serves the business.