In this clip from Data Summit 2026, John O’Brien, Radiant Advisors’ principal advisor and industry analyst, quantifies and demonstrates what differentiates organizations succeeding with agentic AI from organizations continuing to struggle with it and what the data tells us. “A data foundation is not enough for successful AI,” O’Brien said. “Well that goes against everything I’ve heard so far.” He outlined 3 architectural pillars that differentiate successful organizations from struggling ones:

1. Data foundation: This includes data security, semantic definitions, data freshness/latency, unstructured data, and a knowledge graph platform.
2. Trust-oriented: This includes full AI model traceability, automate data trust score, data quality for AI, trust scoring readiness, and metadata/lineage.
3. AI architecture: This includes real-time for AI, fully autonomous agents, multi-agent systems, and agent orchestration.

Full opinion : Why Agentic AI Success Demands More Than a Sound Data Foundation.