Approaches to Data as a Business Discipline for AI

The IT department of any enterprise is integral to implementing and managing the execution of its data objectives, just as the finance department is integral to implementing and managing financial objectives. Few enterprises would allow the finance department complete autonomy to define financial strategies; however, too many enterprises allow the IT department to define data strategies. Treating data as a business discipline—rather than a technical one—is a critical component of delivering competitive advantage through investment in data processing, analytics and artificial intelligence. This can be facilitated by adopting the most appropriate organizational approach, depending on the data activity.

Treating data as a business discipline means data professionals and the IT department remain responsible for managing and processing data, but do so in alignment with a strategic direction defined and led by business leaders to focus on clearly defined organizational goals and key performance indicators. Treating data as a business discipline also means taking an outcome-led approach to data initiatives. I previously explained how “right-to-left” thinking enables enterprises to remain focused on business outcomes even while data professionals are working on the technical capabilities to deliver them, facilitated by new data management products augmented with generative AI.

Treating data as a business discipline also requires being pragmatic when it comes to organizational approaches to data. Almost one-half (49%) of participants in ISG’s 2025 Market Lens Data and AI Program Study agree that data operations should be managed separately from other parts of the IT estate. Few have delivered on that concept, however, as the IT department is the most likely home for all data activities, including AI initiatives (41% of participants) and data insights and business intelligence reporting (31%).

Separating responsibility for data initiatives from the IT department is foundational to treating data as a business discipline, with alternative approaches including the use of a specialist Center of Excellence and the distribution of responsibility to business units. These organizational approaches to data initiatives are not mutually exclusive, and an important result of the ISG Data and AI Program Study is recognition that each has its potential benefits, depending on the focus area of the initiative.

Responsibility for data and analytics can have significant implications, especially for productivity. Almost one-half (49%) of participants at enterprises where the data and insights reporting function is owned by IT say productivity gains from AI and data initiatives are below expectations, compared to only 28% when the data and insights reporting function is owned by a CoE. The concept of an analytics and data CoE is well-established but not very widely adopted. CoEs are responsible for data and insights reporting at almost one-fifth (17%) of participants. The ISG Data and AI Program Study results show that legal and regulatory data compliance and delivering business value from insights reporting are areas where enterprises are more likely to operate above expectations if data initiatives are owned by a CoE.

The distribution of responsibility for data to business units is a growing approach, influenced in part by the desire to deliver data as a product. I assert that by 2027, more than 3 in 5 enterprises will adopt technologies to facilitate the delivery of data as a product as they adapt their cultural and organizational approaches to domain-based data ownership. Data discovery and usability, speed of access to insights and data assets are more likely areas of high achievement when enterprises distribute responsibility for data initiatives to BUs.

Distributed ownership of data has been popularized by the data mesh concept, in which the production and self-service consumption of data products is distributed to business domains, while data governance is federated. The ISG Data and AI Program Study indicates that while responsibility for data and insights reporting is distributed to business units at more than one-quarter (27%) of participants, the distributed approach is much less popular for data governance (11%) and data integration engineering (9%).

Indeed, data integration engineering (61%) and data governance (52%) are far more likely to reside within the IT department, along with data cleaning (56%) and data cataloging (53%). For these initiatives, ownership of data and insights reporting by IT may be advantageous. For example, enterprises with data and insights reporting owned by IT are more likely to be operating above expectations in relation to the delivery of consistent taxonomies and master data across business units. Likewise, enterprises with AI initiatives owned by IT are more likely to be operating above expectations for responsible and ethical use of data and AI.

Good data management is fundamental to enabling enterprise adoption of AI. Organizations that seek to differentiate with data need to treat it as a business discipline with a strategy defined by business leaders outside the IT department. To execute that strategy, I recommend enterprises evaluate the potential benefits of different organizational approaches to data ownership and responsibility and consider the potential benefits of matching the most appropriate organizational approach to the specific data activity.

Regards,

Matt Aslett