AI Hype Detox for Manufacturers and Distributors with Heather Hershey

Episode notes

What happens when you drop an LLM on top of five ERPs and a decade of M&A? Aaron Sheehan and analyst Heather Hershey map the practical path: B2B use cases that work, risks that don’t, and why chunk-by-chunk modernization beats “robot, take the wheel.” Highlights 01:06 – Welcome back and introducing Heather Hershey 03:35 – Defining AI, LLMs, and RAG 09:30 – Why probabilistic AI makes ops teams nervous 11:58 – Is LLM an overkill compared to ‘boring’ machine learning and rule-based systems? 15:19 – The real blocker: fragmented data across ERPs and other systems 19:40 – The strangler pattern: modernize in chunks instead of ripping everything out 21:13 – Why commerce platforms become the orchestration layer for AI/NLP 24:37 – If you had $100K for AI: where to spend it 27:27 – Prisoner’s dilemma: agentic shopping and the disintermediation trap 34:53 – Agentic commerce predictions for B2B 40:30 – Are people replacing Google with LLMs?