Dr.

The CAIPM framework outlines several AI operating models—centralized, decentralized, federated, and hybrid—each suited to different organizational conditions. The key decision factors in this scenario are strict governance requirements, high regulatory exposure, and limited specialized talent .

A Centralized Model is most appropriate when an organization needs strong control, standardization, and consistency across all AI initiatives. In this model, a central team owns AI development, tooling, governance, and deployment, while business units act primarily as consumers of shared capabilities. This ensures that policies are uniformly applied, risks are tightly managed, and scarce expertise is concentrated where it can be most effective.

The scenario explicitly states that business units should consume AI solutions rather than build their own, which is a defining feature of centralization. This approach reduces duplication, enforces compliance, and minimizes variability in how AI systems are developed and used.

Other models are less suitable:

Decentralized models distribute ownership to business units, which conflicts with the need for strict governance.

Federated models allow some autonomy while maintaining coordination, but still require distributed expertise.

Hybrid models combine approaches but are typically used when maturity is higher and talent is more available.

CAIPM emphasizes that organizations early in AI adoption, especially in regulated environments, should adopt centralized structures to establish strong governance and control before scaling.

Therefore, the correct answer is Centralized Model , as it best aligns with the requirements of uniform control and limited expertise.