Architecting the AI-Native Enterprise: Identity as Infrastructure, Technical Debt as Liability, and the Repricing of Enterprise Security

The transition to AI-native enterprise architecture demands a fundamental reconceptualisation of security primitives. This paper argues that identity must be elevated from a supporting service to foundational infrastructure — the control plane through which all AI interactions, both human and machine, are governed. Simultaneously, technical debt can no longer be treated as an operational inconvenience; under regulatory frameworks like DORA and NIS2, it represents material liability that must be quantified, disclosed, and actively managed.

The paper introduces a comprehensive architectural framework that addresses three converging forces: the proliferation of non-human identities driven by AI agents, the accumulation of security-relevant technical debt in legacy systems, and the regulatory-driven repricing of enterprise security from cost centre to balance-sheet item. Case studies from Tier 1 financial institutions illustrate practical implementation of these concepts.