ServiceNow Podcasts Shift Left Everything: How Ontology, Events, and Culture Unlocked Enterprise AI with Nachiket Mehta
Feb 19, 2026
Nachiket Mehta, VP of AI and data engineering who builds ontologies and event-driven data foundations. He recounts mapping real-world operations to system events. He tells the $35M lost-trailer tale and why tiger teams and warehouse visits mattered. He champions shifting data teams left, embedding engineers in domains, and tying work to OKRs to turn messy data into actionable systems.
AI Snips
Chapters
Transcript
Episode notes
Real-World Event Gaps Caused $35M Loss
- Nachiket describes discovering that KPI mismatches were caused by real-world event gaps, not just definition problems.
- He recounts a trailer stuck in a yard that caused $35M in delayed orders and revealed missing event capture.
Ontology Solves Identity, Enables Reliable AI
- Ontology and event-driven context solved identity and meaning problems beyond KPI definitions.
- That context later reduced hallucination when applying generative AI agents.
Map Happy And Unhappy Paths
- Map happy paths and unhappy paths to identify missing events and exceptions in operations.
- Use those mappings to prioritize fixes that directly reduce business losses like lost-in-transit costs.