Investing in Data Governance in the AI Era

Introduction

In 2026, more organizations are putting data governance, a critical framework supporting trust in data, as their critical priority. But more data governance has not necessarily translated to better outcomes. The 2025 DATAVERSITY Trends in Data Management Survey reports that participants list data quality issues and lack of data governance as among the biggest data management challenges their organizations face.

The rapid evolution of AI development and takeaways for adapting to the future play a central role in DATAVERSITY’s new webinar series, The DGIQ Dialogs: Governance in Action. In the first episode, Jimm Johnson, VP of professional development at DGPO, moderated a panel on the “Future of Data Governance and Trends to Watch in 2026.” 

Thought leaders Noah Yao (principal of data strategy and governance at SC&H), Frances Stoor (data and AI governance manager at Jackson), and C. Lwanga Yonke (founder and president of Padouk Consulting, LLC) provided their insights. They emphasized strategic investments that prioritize foundational capabilities, advancing adaptable and purpose-driven frameworks that provide the context of the data with one click. 

As Yao noted, “There is more cultural acceptance and investment in data governance, and we are challenged to do it effectively.”

An AI Wake-Up Call on Data Governance Investment

AI has increased data governance urgency and priorities. Organizations not only need tools, like glossaries or ontologies, so that people can find information, but they also have a “new user,” noted Stoor. Generative AI exemplifies this “digital persona,” which is “ready, eager, and hungry” for data.

Yonke added that “generative AI creates volumes of unstructured data – such as videos, images, and so on – that result in prompts.” He comments that this unstructured data makes up 80% to 90% of an organization’s information assets. 

This unstructured information grows up to four times faster than structured data. Now, as enterprises continue to solve the typical data quality problems with data governance, “AI and generative AI expose a weak data foundation in the typical business,” said Yonke.

Consequently, leaders feel the pressure of governing data for AI with the regular data governance challenges, as Johnson observed. Yao noted that while these requirements have organizations putting the time and money into data governance, they require visible outcomes from this funding – quickly. The resulting stressors lead to misdirected efforts, where companies end up focusing on the wrong activities or stopping too early, creating an illusion of progress without delivering real business value.

Where Data Governance Goes Wrong

According to the panelists, many companies do data governance as a short-term, once-and-done project. Yao explained:

“We grab answers, such as putting formal policies in place, assigning data ownership, or      working on metadata governance. Then we think, I worked on it, and I’m done. I have better data governance. But in a year, nothing valuable comes from these efforts. We need to be more purpose-driven or holistic about the way we think about data governance. What are the initiatives you’re putting in place? What are you trying to achieve?”

Some companies do data governance more comprehensively but over-rely on that one set of frameworks that has been successful. “We have to adapt,” said Stoor. “The pace of change is incredibly fast right now, and we have to prioritize our time on what we assess and define as critical data. We have massive volumes of structured or unstructured data that are consumed quickly. If we’re still stuck on defining critical data elements, maybe we need to move on from there. If our priority does not impact the ability to have the right data for AI, maybe we should table that one and see if some other information sources are fit for purpose.”

Without nimbleness, data governance initiatives become bloated, expending many resources on less important work and outcomes. As a knee-jerk reaction to the whirlwind of data, businesses turn to chasing fads or the latest technology. Yonke cautioned:

“We rush to buy tools without having really understood the processes that we’re trying to automate. Let’s resist that temptation. Most organizations and people are still in the old world, the non-AI-related world. They still need data governance help. We need to focus on those fundamentals.”

Success requires strategic investment in core data governance fundamentals that deliver the highest impact on business objectives – regardless of whether they are AI-related.

Strengthening Context Capabilities

A variety of technologies generate a lot of data that needs governance. Yonke brought up IoT, internet-connected sensors, and computer vision. Add in model context protocol (MCP), and digital personas extract and manage this data depending on the circumstances provided. As Stoor noted, “Context is king.”

Yao emphasized the transformative potential of unified metadata architecture: “Having a consumable link to the metadata layer will change how we do data governance,” both now and in the future. This metadata-driven foundation creates a translatable, usable interface that benefits entire organizations. 

Stoor reinforced this point: “Make sure that your metadata – all the different fundamentals you use to capture your governance decisions – can be leveraged by additional systems.” This foundation becomes critical as organizations must track data lineage, understanding where data originates and how it flows, particularly when AI systems are involved.

Building Collaborative Partnerships

Successfully implementing AI governance requires enterprise-wide collaboration. Yonke argued that while data governance should have significant influence, it shouldn’t dominate the conversation. Instead, he advocated for a partnership model: “AI governance requires collaboration between data governance, IT, business risk, legal, ethics, and so on. If data governance can figure out how to build strong partnerships with all those other disciplines, then data governance and AI governance will have more success.”

This collaborative approach extends beyond organizational structure to execution. The real challenge lies in change management: helping organizations adopt new data governance practices and processes. As Yonke emphasized, “Let’s really become highly skilled in leading organizational changes. We need to be really skilled at collaborating, influencing, and building partnerships. When we do that well, we’ll be more successful.”

Conclusion

Panelists at DATAVERSITY’s inaugural DGIQ Dialogs webinar delivered a clear message: AI has transformed data governance from optional to essential. Organizations are already investing – now they must invest strategically.

Checkbox governance creates an illusion of progress without business value. Success requires three fundamentals: robust metadata architectures, collaborative partnerships across departments, and change management capabilities that drive adoption.

The window for strategic action is narrowing. As AI accelerates, organizations cannot afford to treat their investments in data governance as an afterthought. They need to be purpose-driven and meet business priorities. This approach – and the window for getting ROI – is closing fast.