AI-Native Architecture: Designing Software Around AI as Its Core Engine

Artificial intelligence is shifting from an add-on feature to the central organizing principle of software design, a paradigm known as AI-native architecture. Unlike traditional systems built on hand-written rules and algorithms, AI-native systems embed intelligence across every layer, covering perception, reasoning, planning, optimization, learning, and adaptation. The text describes this as the next stage after mainframe, client-server, web, and cloud-native computing, aiming to optimize system-wide intelligence rather than just infrastructure flexibility. Key characteristics include continuous learning from operational data, model-centric design, and agent-based structures where multiple AI models collaborate, while deterministic control governs safe execution. The chapter is part of an introductory series on software architecture fundamentals.