CIOs Must Redesign Operating Models to Scale Enterprise AI

Enterprise artificial intelligence initiatives are stalling as organizations struggle to bridge the gap between executive mandates and operational reality. While vendors continue to push agentic platforms, the primary constraints remain internal, centered on data quality, governance, and the failure to adapt legacy operating models to the requirements of automated systems.

Data quality remains the most significant barrier to adoption, with 83% of organizations identifying it as their primary challenge, according to Lopez Research. Maribel Lopez, founder and principal analyst at Lopez Research, notes that historical neglect of data hygiene creates compounding technical debt when companies attempt to deploy complex AI models. This systemic issue is further complicated by the fact that only 21% of enterprises report having a mature governance model for AI agents, as highlighted in Deloitte’s 2026 State of Enterprise AI report.

Boris Evelson, vice president and principal analyst at Forrester, argues that most enterprises are not prepared for AI at scale because they treat the technology as a series of disconnected projects. This “use case trap” prevents companies from achieving enterprise-wide impact, often resulting in fragmented priorities and an inability to demonstrate tangible return on investment. The technical challenge is compounded by the fact that AI systems require a level of process clarity that traditional software development cycles often bypass.

Afshean Talasaz, former CIO at Colonial Pipeline, emphasizes that the technology playbooks of the past are insufficient for modern AI deployment. Unlike traditional software built on static user requirements, AI operates within the intersection of data production and consumption, necessitating a deeper, more intentional partnership between IT and business units. This shift requires CIOs to move away from bolting AI onto existing workflows and toward a fundamental redesign of business processes.

Levi Strauss provides a practical example of this transition, where the company established a single source of truth by standardizing over 1,100 operating procedures on top of its SAP environment. Jason Gowans, chief digital and technology officer at Levi Strauss, notes that this foundational work allows the company to feed structured, high-quality data into large language models to automate complex business tasks. The result is a significant reduction in time-to-value, such as accelerating partner onboarding from months to days.

Jabil has similarly focused on driving ownership around data and decision rights to prevent the proliferation of siloed data sources. Chase Christensen, segment CIO at Jabil, highlights that the technical effort of data preparation must be paired with aggressive AI literacy programs to mitigate employee apprehension. By demystifying the technology, organizations can better align their workforce with the new requirements of integrated, industrial-capable AI systems.

The shift toward AI-ready infrastructure requires a fundamental reallocation of capital, with Forrester recommending that roughly 48% of AI spending be directed toward foundational data management and engineering. This investment is necessary to address the reality that most corporate data is currently siloed, poorly governed, and difficult to integrate. As users increasingly generate new forms of data on the periphery of existing systems, the challenge for architects will be to capture and leverage these insights without creating further fragmentation.

Success in this domain is ultimately determined by business-led sponsorship rather than IT-centric delivery models. When business CxOs own the outcomes, AI initiatives are more likely to align with enterprise strategy and avoid the pitfalls of tool-centric implementation. Leadership teams must now articulate a clear vision for how AI will reinvent business experiences, providing the north star necessary for teams to build differentiating capabilities.

The next phase of enterprise AI will be defined by the ability to move from experimentation to integration within core business processes. CIOs should monitor the maturity of their internal governance frameworks and the extent to which their data architecture supports real-time consumption. Future success will depend on the capacity to rewire processes at the core, ensuring that AI systems operate within a well-defined and governed operational environment.