Most enterprises expanding AI initiatives lack the foundational data governance infrastructure needed to ensure reliable, compliant outcomes at scale, according to research published in 2025. The gap between AI adoption and governance maturity is widening across healthcare, financial services, manufacturing, and the public sector-with significant regulatory, financial, and operational consequences.
Background
A 2025 AI Governance Survey by Pacific AI found that 75% of organizations have established AI usage policies, yet only 36% have adopted a formal governance framework. Without a broader framework, many organizations lack consistent roles, controls, monitoring, and enforcement.
A Deloitte survey reported that data governance ranked as the top priority for 51% of Chief Data Officers in 2025, according to research compiled by OvalEdge. IBM reports that approximately 13% of IT budgets were allocated to data strategy in 2025, up from 4% in 2022. Despite increased investment, the operational gap persists.
The EU AI Act entered into force on 1 August 2024 and began phasing in substantive obligations from 2 February 2025. The Act requires companies deploying high-risk AI to document data origins, transformations, and quality metrics, with potential fines reaching USD 39.82 million or 7% of global turnover for non-compliance. These regulatory pressures are forcing a reassessment of governance priorities across regulated industries.
Details
During the opening keynote at the 2025 Gartner Data & Analytics Conference, VP Analysts Gareth Hersche and Carlie Idoine delivered a clear message: AI is only as reliable as the data it is trained on. In a subsequent session, Paul Carey of Bank of New York articulated the core requirement: "To scale AI, you need to know what data you have, where it lives, how it flows, and why it matters."
Organizations lose an average of $12.90 million annually due to poor data quality, according to Gartner research, as reported by Elevate Consulting. Poor data quality, opaque lineage, or weak access controls amplify model bias, erode customer trust, and invite regulatory penalties.
Cross-industry evidence points to distinct governance failure modes. Healthcare teams rely on provenance tracking to record patient data accurately from original collection through subsequent transformations, while financial institutions require provenance to audit transactions and trace them to their origins. This transparency is especially critical in regulated industries such as finance and healthcare, where organizations must justify automated decisions and demonstrate model fairness, security, and integrity.
Only 16% of healthcare organizations have system-wide AI governance frameworks in place, and 81.3% of hospitals have not adopted AI at all, according to Strativera's analysis of peer-reviewed studies. The FDA had authorized approximately 950 AI/ML-enabled medical devices as of 2024-up from just 6 in 2015-and January 2025 guidance on AI-enabled device software lifecycle management emphasizes data quality, algorithm transparency, and change management as foundational requirements.
On governance architecture, KPMG's 2025 analysis identifies an ideal model integrating AI and data governance under a single umbrella-one that enables complete transparency, creates enforceable policies and standards, eliminates duplicate data sets, and uses data and AI use cases to deliver tangible value. Forward-thinking enterprises are aligning governance models with global standards such as the NIST AI Risk Management Framework or ISO/IEC 42001:2023, which provides a structured approach for AI management systems.1What is AI-Powered Data Lineage? A Complete Guide | Devoteam
Practitioners are moving toward automated lineage as the operational backbone of scalable governance. AI-ready data lineage is defined as lineage metadata that is complete, granular, continuously updated, and accessible to both humans and AI systems-providing the context AI agents need to understand, trace, and act on data across the enterprise ecosystem. Machine learning-powered lineage systems can handle massive, complex data pipelines far beyond what manual documentation allows, scaling to process large data volumes across multi-cloud environments.
Organizations with mature AI governance focus strategically on fewer high-priority initiatives and achieve more than twice the ROI of other companies, according to BCG research published in 2024.
Outlook
Enterprise practitioners and analysts point toward dynamic, federated governance models that embed oversight directly into data and analytics workflows-replacing static policy documents with automated, continuously enforced controls. A 2025 leadership snapshot from Evanta reports 65% of data leaders are investing in AI, while 44% invest in data governance and 41% in data quality-a convergence analysts say will accelerate as regulatory enforcement intensifies under the EU AI Act's phased timeline through 2026. For CIOs and CDOs, the operational imperative is clear: governance infrastructure must precede, not follow, the scaling of AI into production systems.
