While predicting the future may be impossible (so far), analyzing trends and learning from industry leaders can help us get pretty close. In an effort to better understand where data governance is heading, we spoke with top executives from IT, healthcare, and finance to hear their thoughts on the biggest trends, key challenges, and what insights they would recommend.
Without spoiling all the insights, the road ahead for governance will require adaptability, innovation, and a forward-thinking approach. This blog is a collection of those insights, but for the full trendbook, we recommend downloading the PDF.Â
Lastly, we partnered with our friends at Fivetran to help bring these insights to life in a meaningful way. With that, let’s get into the governance trends for data leaders!
No problem! Just click this button and fill out the form to download it.
(You can read the full guide without giving us your email — keep scrolling!)
What is the Impact of Data Governance on GenAI?
Data governance plays a critical role in the successful implementation of Generative AI (GenAI) and large language models (LLM), with 86.7% of respondents rating it as highly impactful.Â
It serves as a vital protective measure, ensuring proper data access while managing risks like data breaches and unauthorized use. Strong data governance also lays the foundation for better model performance, cost efficiency, and improved data quality, which directly contributes to regulatory compliance and more secure AI systems.
Quotes
It's extremely important because many of the Gen AI and LLM applications take an unstructured data approach, meaning many of the tools require you to give the tools full access to your data in an unrestricted way and let it crawl and parse it completely.
The impact on data governance due to GenAI/LLM is that these technologies can spot trends much faster than humans or other applications. They can identify where risks are and what to avoid.
Challenges and Considerations
Balancing data access and protection is essential as GenAI tools require broad access while still adhering to governance policies. Some respondents feel current governance practices may not fully align with the capabilities of LLMs and GenAI, and there’s a tension between the need for speed in implementing AI solutions and the time needed for proper governance.Â
Organizations also need a better understanding of how LLMs are trained, especially with external vendors or public cloud environments. In sectors like legal services, safeguarding client data from being used in public apps or external training models is critical.
Quotes
Data governance is going to play a large role in what data can go into an LLM. Right now, many external vendors have strict language around which content can go into an LLM.
It will be increasingly important for organizations to understand how LLMs are trained -- whether on the company's own data or paired with others. Additionally, how your data might be used across vendor LLMs when used in public cloud scenarios will be important to ensure that valuable IP or PII does not leave your platform.
Recommendations for Data Leaders
Align Data and AI Governance: Ensure that data governance strategies are aligned with AI governance to support successful GenAI implementations.
Implement Robust Data Controls: Focus on establishing standardized data controls to improve data quality, enhance model performance, and ensure compliance.
Educate on GenAI Capabilities: Invest in understanding the full capabilities and potential risks of GenAI and LLM applications.
Balance Innovation and Protection: Develop strategies that allow for rapid innovation with GenAI while maintaining strong data protection measures.
Monitor and Adapt: Continuously assess the impact of GenAI on data governance practices and be prepared to adapt policies as technologies evolve.
Quotes
It's important to ensure that the right and proper data is being input and used for GenAI and LLMs, as well as ensure that proper data management is in place to adhere to client rules regarding GenAI use. Data governance is the only way to ensure those requirements are met.
Survey respondents noted improved data quality and compliance and risk management as the top two outcomes for organizations with a focus on more standardized data controls when working to implement GenAI and LLM applications.
Data Governance and Modern Data Management
AI and machine learning (AI/ML) applications emerged as the leading trend in data management, significantly shaping organizations’ data platform strategies. As data volumes grow, scalable solutions like data mesh and data fabric architectures are becoming more widespread due to their flexibility alongside complex organizational structures.Â
Additionally, more organizations are continuing to migrate their data to cloud platforms to meet demands for scalability, security, and cost-efficiency, making cloud adoption another important trend in data management.
Survey respondents indicated faster time-to-insight, increased scalability and flexibility, and reduced bottlenecks and dependencies on centralized data teams as the top three effects on the agility of data teams following the implementation of a data mesh architecture.
Quotes
GenAI and LLM will impact data platforms as they need a bigger amount of data to better train the models.
Moving data to the cloud is generally a move towards centralization, it doesn't have to mean moving towards a singular record of truth or MDM model, but moving data from many proprietary systems into a generalized cloud data solution at least lets individuals in the organization start the process of information to data transition.
Survey respondents specified easier risk management and more data access to personnel as the top two benefits organizations can expect from moving data into a cloud platform.
Challenges and Considerations
Implementing data management in the age of AI comes with challenges, such as ensuring proper security and compliance while also needing high-quality, comprehensive data for accurate results. Organizations must balance agility with strong controls and standards, especially when managing data across multiple cloud platforms.
Developing skills is also crucial, as teams need education on new architectures and technologies, along with fostering collaboration across different areas. Additionally, managing costs is a key consideration, particularly in cloud environments where infrastructure expenses can easily escalate if not controlled.
Quotes
There are so many companies rushing into the AI/ML space and producing generative AI and LLM-based tooling or products, I fear many organizations will make use of poorly implemented and secured systems.
Organizations have spent the past decade accumulating, maintaining, and securing data lakes/warehouses/fabrics that will now be expected to drive AI/LLM use cases. Many organizations may be surprised in the quality of results delivered -- either because their data was not complete (in describing the full business scenario) or being of sufficient volume to drive the proper trending needed.
Recommendations for Data Leaders
Develop a clear roadmap for AI/ML integration in data management processes
Carefully plan cloud migration strategies, addressing security and governance concerns
Invest in team education and skill development for new data architectures
Implement robust cost management and monitoring for cloud data platforms
Focus on improving data quality and completeness to support AI/ML initiatives
Enhance metadata tagging and management to improve data discoverability and usability
Quotes
Data mesh architecture implementation is the long-term approach, with a roadmap to implement domain-based mesh architecture. Iterative development with the MVP for key business domains, e.g., finance, maybe a good starting point.
Organizations will focus more on metadata tagging of existing and new content in the coming years.
No problem! Just click this button and fill out the form to download it.
(You can read the full guide without giving us your email — keep scrolling!)
Data Governance Maturity
For many organizations, data governance is still in the early stages, focusing on defining policies and processes. However, real success requires more than just a plan—it needs strong executive support, clear rules, and regular involvement from stakeholders.
Looking ahead, data governance will focus on real-time management, data privacy, ethics, and easier access to data for everyone. Self-service analytics tools and technologies like data lineage and provenance tracking will also play a key role in making data management more transparent.
Quotes
I think many organizations practice data governance at a very high level, mostly at the policy and process definitional level. Many organizations that have been a part of that try to institute data governance end up slowing things down due to not thinking through the compliance and auditing pieces fully in order to ensure a lightweight impact is felt throughout the organization, data governance programs can crush productivity through restricting access to relevant information in a timely manner.
The organizations are mostly early in their journey with areas like data governance frameworks established; however, the execution of the frameworks, ownership, and understanding of data literacy, especially within the business, is fairly low maturity. The technology for metadata management, data quality management, etc., is fairly advanced. However, implementation in a large complex environment is difficult due to investment challenges and buy-in from the business.
Survey respondents selected data encryption as the most observed practice organizations are currently using to maintain data security.Â
Challenges and Considerations
Organizations are still figuring out how to effectively implement data governance strategies, facing challenges like employee turnover, lack of compliance among business partners, and difficulty in aligning stakeholders.
There’s mixed agreement on whether clear patterns exist for identifying and protecting sensitive PII data. While some patterns are in place, they aren’t consistently applied, especially in organizations with lower regulatory demands. The complexity of data and the challenges of marking PII during intake add to the difficulty.
To ensure data quality, platforms need consistent, automated processes with continuous testing and validation. Systems should include alerts to flag any changes or anomalies that could affect data integrity.
Quotes
Quality is a never-ending exercise of testing, validating, and alerting to changes or patterns that can cause your data to be impacted. Security can feel slightly similar, but with proper tooling (like a Vanta, SecureFrame, or Drata), most of your structural or architectural controls can be handled. Compliance is the final, and hardest, mile as ensuring the very access that drives value within your business does not make it too easy to leak data either from lightly controlled websites to oversharing of information.
The source of the data is the starting point for everything. It must be a trusted source that has been vetted and randomly tested upon intake. Afterwards, periodic data audits must occur to ensure the accuracy of the data.
Recommendations for Data Leaders
Invest in metadata and data lineage systems
Foster a culture of data accountability
Focus on data privacy and ethical use
Automate security processes where possible
Continuously educate users on data security best practices
Regularly audit and assess data quality and security measures
Survey respondents marked the adoption of advanced data governance technologies as the most anticipated change to how organizations will handle the introduction of more data sources, data types, consumers, and downstream use cases in the coming years.
Quotes
Data governance assures that proper policy, procedure, personnel, communications, and enforcement is in place to assess the data estate and implement appropriate measures to ensure that data is created, maintained, secured, and disposed according to regulatory, client obligations, and best practices over time.
Using industry standard frameworks and models, creating a culture of data accountability, and ensuring data governance is part of an overall data strategy.
Conclusion
Data governance is constantly evolving, especially with the rise of generative AI technologies. As businesses adopt these innovations, one thing is certain: change will be a constant in managing data and AI/ML applications in the coming years.Â
Business leaders must be ready to adapt and guide their teams with agility, embracing the ongoing shifts in technology and regulation. The key to success will be staying flexible, proactive, and responsive to these changes.
