Building the Agentic Enterprise: A Practical Field Guide to Designing, Deploying, and Operating Agentic AI Systems on Google Cloud
暫譯: 打造 Agentic 企業：在 Google Cloud 上設計、部署與運作 Agentic AI 系統的實務指南

Build production-ready agentic AI systems that move beyond pilots with governed architecture, multi-agent orchestration, AgentOps, EvalOps, and Google Cloud.

Key Features:

- Bridge the Day 2 production gap with secure, governed, cost-aware agent architecture patterns

- Engineer around the model with context, Harness Engineering, AgentOps, and EvalOps

- Scale human-agent delivery with the AI Squad Model beyond the Supervision Ceiling

- Purchase of the print or Kindle book includes a free PDF eBook

Book Description:

AI can now produce credible demos, prototypes, and proof-of-value concepts faster and more cheaply than most enterprises can absorb. Yet far fewer ever reach production. Building the Agentic Enterprise is a practical field guide to closing that gap: identifying which ideas are worth pursuing and designing the architecture, controls, and operating model needed to make agentic systems reliable, governable, and repeatable.

Drawing on two decades of experience across cloud, microservices, and AI, and three years of taking agentic systems into real-world delivery, the book shows why the model is only one part of the challenge. Using Google Cloud and ADK as the practical lens, a continuous case study grounds the patterns, trade-offs, and decisions in a realistic production setting.

You will learn how to design agent architectures, orchestration, context, memory, security, observability, governance, and evaluation, while building an AI Squad Model where human experts supervise specialized agents within clear boundaries. The book then expands to the platform and operating model required to scale safely across an organization.

By the end, you will be able to judge where agents create real business value, choose the right level of autonomy, and build an agentic capability your organization can trust, operate, govern, and scale.

What You Will Learn:

- Distinguish proof of value from operable, governable agent systems

- Choose agent architectures while accounting for the Coordination Tax

- Design orchestration around the Three Cs for reliable outcomes

- Engineer governed context, memory, and harnesses for reuse at scale

- Define Agency Scope to expand autonomy deliberately and safely

- Separate AgentOps platform health from EvalOps decision quality

- Persist decisions, state, and evidence for audit and evaluation

- Scale AI Squads beyond the Supervision Ceiling with human accountability

Who this book is for:

Enterprise architects, solution architects, technical directors, principal engineers, applied AI engineers, cloud consultants, AI platform teams, and engineering leaders who need to take agentic AI from pilots to governed production systems. A general understanding of modern software systems, cloud platforms, and APIs is expected; Python familiarity is helpful but not required.

Table of Contents

- Why Agentic Systems Fail to Go Live

- The Agentic AI Foundation

- Single-Agent and Multi-Agent Systems

- Core Architectural Patterns for Agentic Systems

- Multi-Agent Orchestration - From AI Pioneers to the Agentic Enterprise

- Context, Memory, and Knowledge

- Security, Identity, and Guardrails

- Building and Deploying Agentic Workflows

- State, Memory, and Decision Persistence

- AgentOps and EvalOps

- Platform Engineering for Agentic Systems