Agentic AI Outpaces Governance, Experts Call for Dynamic Data Fabrics

TL;DR: Enterprises are deploying agentic AI faster than they can govern it, prompting experts to recommend centralized lakehouses, data mesh architectures, and dynamic data fabrics for control.

Context Enterprises are moving from AI experiments to production‑grade agentic systems that can act autonomously across business processes. The speed of adoption has created a gap: organizations lack the data governance, observability, and identity frameworks needed to keep these agents secure and accountable.

Key Facts Jena Zangs, chief data and AI officer at the University of St. Thomas, says a centralized data lakehouse combined with a data mesh and metadata tagging delivers governability. The approach limits agents to specific domains, preventing unrestricted database access. Pablo Ballarin, co‑founder of Balusian S.L. and ISACA Emerging Trends Working Group member, warns that siloed data stores and static warehouses cannot support secure, governable AI agents. He advocates for dynamic, entity‑centric governed data fabric architectures that treat data as interconnected, policy‑driven entities.

Enterprises are rapidly adopting agentic AI, but governance controls are not keeping pace. The mismatch threatens data privacy, compliance, and operational stability as autonomous agents make decisions without human oversight.

What It Means Organizations must shift from legacy data warehouses to flexible data fabrics that embed governance at the entity level. Centralized lakehouses provide a single source of truth, while data mesh layers distribute responsibility across domains, and metadata tags encode access rules. Together, these components create a controllable environment where agents can retrieve only the data they need.

Adopting such architectures will require investment in data cataloging tools, policy engines, and real‑time monitoring. Companies that fail to implement dynamic governance risk regulatory penalties and loss of stakeholder trust.

Looking Ahead Watch for industry standards on entity‑centric data fabrics and emerging tools that automate policy enforcement for autonomous AI agents.