Operating Models & Enterprise Architecture in the Age of AI

Overview

This white paper provides a strategic and architectural analysis of how an organization’s fundamental operating model—Diversification, Coordination, Replication, or Unification— determines the correct design of its AI Target Operating Model (AI TOM). It reframes AI architecture not as a technology-first initiative, but as a direct expression of how the enterprise chooses to operate.

The paper also introduces the AI‑First Enterprise Architecture Stack, a four‑layer model covering AI intelligence, semantic meaning, data products, and AI‑ready data management. It explains how each layer should be centralized or federated depending on the operating model.

What’s Inside

• The MIT CISR Operating Model Framework
• Why Operating Model Choice Must Precede AI Architecture
• The AI‑First Enterprise Architecture Stack (4‑Layer Model)
• AI TOM Implications for All Four Operating Models
• Cross‑Framework Synthesis with McKinsey & MIT CISR AI Research
• Strategic Recommendations for Chief Enterprise Architects

Executive Summary

Artificial intelligence is becoming the primary mechanism through which enterprises sense market signals, make decisions, automate work, and create competitive differentiation. This transformation requires a corresponding evolution in operating model design and enterprise architecture.

The paper argues that there is no universally correct AI Target Operating Model. The right AI TOM is the one that is coherent with—and in some cases deliberately designed to evolve—the organization’s operating model. Misalignment between operating model and AI architecture is one of the most common and costly failure patterns in enterprise AI programs.