Data Governance, AI and Depth of Thought

Written By Alana Moran

Having a crisp mental model around a problem, being able to break it down into steps that are tractable, perfect first-principle thinking, sometimes being prepared (and able to) debate a stubborn AI — these are the skills that will make a great engineer in the future, and likely the same consideration applies to many job categories” (Argenti, 2024).

The article Why Engineers Should Study Philosophy validates the importance of having a crisp mental model around a problem, which is how companies should approach the implementation of Generative AI — on the basis of crisp models, supported by quality, governed data. The cornerstone of this process is Data Governance, a framework that companies must implement and maintain. This has been a challenge for organizations even before the prevalence of AI, but the issue has worsened rather than improved.

The Reality

·       In a recent global study of more than 1,300 tech and data executives, just 18% of companies report being fully ready for AI deployment, meaning their data is fully accessible and unified (Curry, 2024).

·       A lack of trust in AI, regulatory challenges, and governance issues are all hindering successful AI deployment, causing many projects to get stuck in the planning stages (Qlik, 2024).

Focus on Data Governance and Data Quality

·       A Harvard Business Review survey found that 46% of data leadersidentified data quality as the greatest challenge to realizing generative AI’s potential in their organizations (HBR, 2024).

·       Data governance tasks, such as data cleaning, integration, and establishing metadata, are foundational for creating a reliable data environment for AI. Experts highlight that data governance must focus on specific business domains where generative AI will first be implemented, such as customer service, marketing, and R&D (Rupanagunta, Sayed, & Arni, 2024).

Action Items

To effectively implement a data governance strategy and harness generative AI, consider these steps as outlined by Mathematica (2024):

Start with Business Drivers

·       Determine the organization’s business priorities — think KPIs.

·       Align KPIs with the organization’s mission and vision.

Identify Key Stakeholders

·       Identify those collecting, curating, and using data.

·       Formalize data-related roles to enhance efficiency.

Document Data Locations and Access

·       Assess who has access to data and define their permissions.

·       Conduct an inventory to ensure appropriate access control.

Create an Enterprise Glossary

·       Establish a common language for data use across the organization.

To leverage generative AI effectively, organizations must start with the “why” — understanding their objectives and aligning their strategy with business goals. Quality and governance of data must remain at the forefront, supported by a philosophical mindset and a healthy dose of skepticism and common sense throughout the journey.

References

·       Argenti, M. (2024). Why Engineers Should Study Philosophy. Harvard Business Review. Retrieved from HBR

·       Curry, R. (2024). The Hardest Part of Deploying Gen AI for Most Companies Is Having Data That’s Ready. NBC News.

·       Mathematica. (2024). Data Governance Is Critical to Getting AI Right. Retrieved from Mathematica

·       Qlik. (2024). 1 in 5 UK Businesses Halt AI Projects Over Trust and Governance Concerns. Retrieved from Qlik

·       Rupanagunta, K., Sayed, I., & Arni, R. (2024). Is Your Company’s Data Ready for Generative AI? Harvard Business Review. Retrieved from HBR

 

Alana Moran