Part 2 — Designing Agentic AI Infrastructure

Overview

Part 1 established what modern AI can do. Part 2 examines what changes when AI systems are given goals, tools, memory, and constrained autonomy.

The shift from generative to agentic AI is not simply a matter of better models. It requires a different architecture, a different set of design decisions, and a fundamentally different relationship between human operators and the systems they deploy. An agentic system does not merely produce a response; it can be designed to plan steps, use tools, observe results, and adjust its next action within boundaries set by humans and organisations. That loop changes everything: the failure modes, the cost structure, the governance requirements, and the competitive dynamics of the market building this infrastructure.

Part 2 maps this shift across four dimensions. First, the core architectural question: when does a single agent suffice, and when does a system of agents become necessary? Second, the platform landscape: who is building the infrastructure that makes agentic deployment possible, and how do those choices constrain or expand your options? Third, the emerging reality of ambient AI — agents that run continuously rather than on demand, and the profound implications that follow. And fourth, the build-vs-buy decision that every organisation must make under conditions of uncertainty, rapid platform change, and high operational stakes.

---

Chapters in This Part

Chapter	Title	Theme
5	One Agent or Many? Designing for Scale and Complexity	Agent architecture
6	The Platform Wars: Who Is Building the Agentic Infrastructure	Platform and infrastructure landscape
7	Always-On AI: The Era of Ambient Intelligence	Continuous agents
8	The Build-vs-Buy Decision in an Agentic World	Build-vs-buy strategy

---

Readers with a primary interest in technical integration may wish to read Chapter 5 alongside Chapter 11 in Part 3.