Generative AI Drives Data Governance

Just as with almost every aspect of an organization, data governance and analytics are significantly impacted by generative AI. This technology emerges as a fundamental tool for addressing historical governance challenges and anticipating future scenarios.

Therefore, let's start with a brief overview of the "pain points" associated with data governance. Although the topic of data governance has been in the market for several decades, since the emergence of the business intelligence concept in the 1990s, it has not fully consolidated, especially due to the rapid growth in data generation volume. Many companies adopted highly audit-focused governance models, which focused on controlling and creating obstacles rather than becoming allies and enablers.

Meanwhile, various organizations have sought to implement best practices. Still, their projects remained "on paper": data governance exists formally, but it is limited to a set of static role and function definitions, disconnected from the company’s technological dynamics and organizational context.

As mentioned, generative AI arrives to transform data governance from two perspectives. The first is the use of this technology to enhance what we already know. For example, in areas as seemingly simple as identifying data sources—which can be highly complex in large systems with multiple branches—, anticipating data quality issues, filling in missing data, or generating synthetic data to improve model training when the available real data is insufficient.

Another revolutionary concept brought by generative AI is metadata management: it allows the management of data that describes other information assets within the organization, improving usability and the ability to find the right data at the right time while reducing inefficiencies and risks.

The second perspective where generative AI will play a crucial role relates to aspects we are just beginning to explore: new players joining the data and analytics ecosystem.

Organizations must broaden their perspective to adapt to projects with methodologies and characteristics different from traditional ones, implement governance for unstructured data (such as audio, video, text, and images), understand how new data models work, ensure data sampling quality for training these models, and guarantee that data is reliable and bias-free, among other challenges inherent to evolving data ecosystems.

In conclusion, generative AI is a strategic asset for shaping a future where data is managed with security and efficiency, driving the highest possible value for businesses.

—Lina Maria Bello Galindo, Data & Analytics Director at NTT DATA