Many organizations are running some kind of data governance program and some kind of AI governance program. Very few run both well. The 2025 DATAVERSITY Trends in Data Management survey found that only 4% of organizations have achieved high maturity in data governance and AI governance at the same time.
One reason the number is so low is a common misconception that AI governance is data governance applied to models. The two disciplines share structural DNA, including decision rights, risk oversight, and accountability, but they govern different things. Data governance manages data as an enterprise asset. AI governance manages AI solutions across their full lifecycle, including how those solutions reason, decide, and act.
That said, if you have a working data governance program, you are on the right path to establishing AI governance. The organizations that reach the 4% evolve data governance in service to AI and stand up AI governance as its own discipline that runs in parallel, connected to the foundation they already have.
Confidence is Running Ahead of Readiness
Part of the problem is that organizations believe they are further along than they are. In the 2026 State of Data Integrity and AI Readiness research from Precisely and Drexel University's LeBow College of Business, 87% of data and analytics leaders said they have the infrastructure, skills, and data readiness AI requires. Between 41% and 43% of those same leaders named those exact areas as their biggest obstacles.
The gap is costly. Gartner projects that through 2026, 60% of AI initiatives will be abandoned due to insufficient data quality.
Data governance is what separates the organizations that close the gap. In the same Precisely and Drexel research, 71% of organizations with governance programs reported high trust in their data, compared with 50% of those without one. Leaders also credited governance with improving AI readiness (42%) and the quality of AI outcomes (39%). Governance also remains a stated priority. In Info-Tech Research Group's Future of IT survey, 40.9% of leaders listed improving data governance as one of their top data priorities for 2026.
What Your Data Governance Program Gives You
A mature data governance program puts several things in place that AI governance depends on.
Ownership and Stewardship
Your data owners and stewards already know who is accountable for which data, including the data AI consumes. AI governance adds a different kind of accountability: who can approve, deploy, monitor, or retire an AI model or agent. Your stewards answer for the data. AI governance needs people who answer for the solution.
Policies and Standards
Your classification, privacy, and quality policies already govern the data AI systems use, and they stay in place. AI governance still needs its own policies and standards for questions data policy was never designed to answer, such as risk tiering, approval paths, and review of AI outputs.
Metadata and Semantic Layer
Your catalog, glossary, and lineage are the evidence auditors will ask for when they want to know what data a model used, which we covered in our post on metadata as the evidence layer for AI governance. Your business definitions also give AI systems the context to reason accurately. AI governance adds metadata types with no equivalent in data governance, such as model provenance, bias assessments, and explainability factors.
Decision Rights and Structure
Your governance council already knows how to make and escalate decisions across business units. That experience carries over to a separate AI stewardship committee that works alongside data stewardship, with both reporting into a unified Data, Analytics and AI Leadership Committee.
When you have a strong foundational data governance program, your AI governance can come together quicker.
How to Build From Your Data Governance Foundation
Evolve data governance in service to AI. As AI consumption grows, data governance expands from critical data elements to datasets, data sources, and AI-consumable data products, and from structured data to semi-structured and unstructured data. Your semantic layer grows from a business glossary into context AI systems can reason with.
Stand up AI governance as its own discipline, with distinct strategy, policies, standards, and stewardship. Start by cataloging the AI already in use, including AI features vendors have added to applications you own. Invite the teams running AI in production to the table early. Your data science team likely has guidelines that can jump-start the program.
Establish decision stewardship as a new function. When an AI system makes or recommends a decision, someone needs to examine whether that decision is explainable, ethically sound, and aligned with your values. Decision stewardship is distinct from data stewardship and will not emerge on its own. It has to be deliberately created.
Connect the two through one roadmap. Distinct programs still need shared milestones and coordinated priorities. At FSFP, the Data Decision Framework is the convergence point where both programs align. The goal is integration without redundancy: identify what the programs share, and govern what is new with the rigor AI's risk profile requires.
Why AI Governance Has to go Further
Data governance was built to manage passive assets. AI agents reason, decide, and act. An agent recommending a credit decision or executing a business process behaves more like a co-worker than a database. Data governance asks whether data is accurate and accessible. AI governance has to ask whether a decision is contextually valid, ethically grounded, explainable, and aligned with your values. Your data governance program gets you to the starting line faster. AI governance takes you the rest of the way.
Where to begin
Take a look at your data governance program. If the business sees it as slow or isolated, that perception will follow AI governance. If it is seen as a service that helps teams move faster, you have a strong base to build on. From there, map where AI is already in use, identify the data and decisions it touches, and define what each program owns.
For a structured view of the full set of capabilities, FSFP's AI Governance Playbook outlines seven building blocks for enterprise-ready AI.
