Part I Overview and Platform Perspective¶
4 chapters · Serving as the book's introduction, responsible for establishing a common language, platform viewpoint, and reading roadmap
The task of Part I is not to dive immediately into a particular framework, code segment, or product, but first to answer four more fundamental questions:
- What exactly is an Agent, and how does it differ from RAG, Copilot, or Workflow?
- Why do enterprises truly need a platform instead of a collection of disconnected intelligent projects?
- What is an AI-native business system, and how does it differ from simply "adding AI features to existing systems"?
- Why are the subsequent 55 chapters organized in the current order, and how should they be read?
These four chapters jointly serve as the "foundational overview" of the volume. They do not rush into specific implementations but rather pursue three objectives:
- First, clarify concepts and boundaries;
- Then, raise the discussion to the enterprise and business system perspectives;
- Finally, provide the overall book map and recommended reading path.
Relationship Among the Four Chapters¶
| Chapter | Question Answered | Role for Next Chapter |
|---|---|---|
| Chapter 1 The Nature of Agents: From Conversational Assistants to Task Execution Systems | What qualifies as an Agent, and why enterprises enter a new problem domain once systems "do work" | Sets the stage for Chapter 2's "why platform boundaries emerge" |
| Chapter 2 Boundaries of Enterprise-Level Agent Platforms | What exactly enterprises are building as a platform, and the distinct roles of platform, application, and framework | Sets the stage for Chapter 3's "which business systems platforms ultimately serve" |
| Chapter 3 AI-Native Business Systems: How Agents Reshape Enterprise Software | Why AI-native is "adding another chat box," and a fundamental change in business system form | Sets the stage for Chapter 4's "why a full book map and architectural panorama are needed" |
| Chapter 4 Complete Book Map: Platform Reference Architecture and Reading Path | The architecture map of the whole book, chapter dependencies, reader paths, and platform building roadmap | Leads readers into the deep dive of the subsequent 55 chapters |
Recommended Reading Approach for This Section¶
If you are encountering enterprise-level Agent platforms for the first time, we recommend reading the four chapters in order, as they form a progressive argument chain. If you already have some project experience, you can skim selectively as follows:
| Your Current Top Question | Recommended First Chapter to Read |
|---|---|
| Which requirements are suitable for building Agents | Chapter 1 |
| The team is debating "What exactly is a platform?" | Chapter 2 |
| Want to assess which business systems are fit for AI-native transformation | Chapter 3 |
| Want a quick overview of the book structure and roadmap | Chapter 4 |
After completing Part I, readers should have developed at least these three capabilities:
- The ability to distinguish different types of large model systems;
- The ability to view enterprise AI projects from a platform perspective;
- The understanding of why each technical topic in the following chapters exists.
Armed with these three capabilities as you enter Part II, the upcoming chapters on models, data, knowledge, runtime, evaluation, security, and frontend will no longer be isolated facts but will gradually assemble into a complete engineering blueprint for enterprise-level Agent platforms.