By
·
1 minute read
McKinsey recently shared research on how organizations structure AI adoption using hub (centralized) versus spoke (decentralized) models. The key insight is that AI adoption is not centralized as a whole—different dimensions of AI operate very differently.
The original McKinsey analysis is available here.
Risk & compliance and data governance are the most centralized elements. In contrast, AI adoption and change management are largely owned by business units. Strategy, roadmaps, technology, and talent typically sit in hybrid models that combine central direction with decentralized execution.
What this analysis does not address is how organizations operationalize this split in practice.
In reality, this often results in a two-layer operating model: centralized, governance-focused platforms at the hub, combined with decentralized delivery and experimentation in the spokes.
On the hub side, we see platforms focused on AI governance, risk, and lifecycle oversight, such as Credo AI, LatticeFlow, and Lumenova AI. These are complemented by more specialized solutions like Lakera for LLM and GenAI security controls, and Aparavi or Concentric AI for data discovery, classification, and access control for AI-ready data.
On the execution side, platforms such as Airia or Sema4.ai aim to empower business units with no-code / low-code tools for building agents and workflows—operating within centrally defined guardrails but close to business value creation.
The real challenge, therefore, is not hub versus spoke.
It is deciding where to draw the boundary between centralized control and decentralized value creation—and how to enforce that boundary in a way that preserves speed, trust, and compliance without slowing adoption.
At AyDEO, we approach this challenge from a different angle.
Rather than treating AI systems as tools or standalone agents, we treat them as Digital Employees. This framing is not semantic debate; it has very practical consequences. By viewing AI as part of the workforce, organizations can apply familiar governance structures they already use for human employees: defined roles, scopes of responsibility, access rights, oversight, performance tracking, and accountability.
This allows AI to plug naturally into existing enterprise architecture, security models, and operating structures, without requiring organizations to throw away what they already have in place. Digital Employees operate under centralized governance by design, while being deployed where value is created—inside business units and end-to-end processes.
In that sense, AyDEO does not replace the hub-and-spoke model. It makes it operational—by providing a concrete abstraction that connects centralized governance with decentralized execution in a way enterprises already understand and trust.