Redefining Enterprise Architecture In The Age Of Artificial Intelligence: From Static Governance To Adaptive, Ai-Augmented Practice

Abstract

The Enterprise Architecture (EA) practice faces a fundamental inflection point as artificial intelligence systems become both objects requiring architectural governance and transformative tools for accelerating EA work itself. Traditional EA frameworks, including TOGAF, the Zachman Framework, and FEAF, were designed to govern deterministic, rule-based information systems and exhibit critical structural gaps when applied to non-deterministic, continuously learning AI systems characterized by emergent behaviors, opaque decision logic, and dynamic capability drift. This paper investigates two interconnected dimensions of this challenge: first, the inadequacies of current EA frameworks as governance instruments for AI systems, quantified through a governance coverage gap analysis; and second, the emerging application of AI as an augmentation tool for EA practice itself, through natural language processing, graph neural networks, and predictive impact modeling. This review article synthesized previous empirical, framework, and practitioner literature from 19 peer-reviewed articles published between 1987 and 2026 to characterize the transition from pre-2.0 EA to EA 2.0 and establish four computable indicators: the Governance Coverage Gap (GCG), AI Augmentation Capability Multiplier (AACM), Digital Transformation Success Factor (DTSF), and EA Technology Readiness Index (ETRI). Results show that TOGAF 10 incorporates approximately 33% of the identified AI governance requirements and that AI-powered EA tools increase the throughput of modeling EA artifacts by a factor of 5.1. The paper concludes with a capability framework for transitioning to EA 2.0 and identifies priority research gaps.