Governance-First AI: Board Authority. Architectural Precision. Governed AI at Enterprise Scale.
暫譯: 治理優先的 AI：董事會權責、架構精準度與企業級治理 AI
Hossain, Mohammad Maruf, Ponnusamy, Ahilan Ayyachamy Nadar
- 出版商: Apress
- 出版日期: 2026-10-28
- 售價: $2,400
- 貴賓價: 9.5 折 $2,280
- 語言: 英文
- 頁數: 542
- 裝訂: Quality Paper - also called trade paper
- ISBN: 9798868831171
- ISBN-13: 9798868831171
-
相關分類:
Large language model、Domain-Driven Design
尚未上市，無法訂購
相關主題
商品描述
As Generative AI shifts from early excitement to real-world implementation, organizations face a critical crossroads. Executives demand rapid returns on investment, yet deployment is frequently stalled by risk, compliance, and governance barriers, and the rise of unmanaged Shadow AI systems has elevated the stakes. Governance-First AI confronts this tension head-on, offering a practical, execution-ready guide for leaders seeking to move beyond hype and into secure, scalable adoption.
This book bridges the gap with a unified, governance-first approach that enables enterprises to operationalise Generative AI confidently and responsibly. It confronts a landmark MIT NANDA finding -- that 95% of enterprise generative AI initiatives deliver no measurable return -- and traces the failure to its root: a systemic disconnect between solution architecture and enterprise governance.The authors present a cohesive framework for achieving Certified Operational Velocity, helping organisations evolve from experimental pilots to compliant, production-grade systems. At its core, the book translates stringent governance principles--including platform engineering and platform-as-a-product thinking--into fifteen concrete architectural patterns. These blueprints address the full spectrum of enterprise AI challenges, from retrieval-augmented generation (RAG) deployment to context management, legacy system integration, and the auditable controls required for autonomous Agentic AI systems. Each chapter delivers step-by-step guidance for building secure, measurable, and future-proof AI capabilities.
By introducing a unified governance-first architecture and pairing it with repeatable, production-ready use cases, Governance-First AI becomes the missing guide for unlocking AI's true impact. It equips organizations to achieve transformative productivity gains--without compromising compliance--while preparing for the era of safe, autonomous AI.
What you will learn:
- Implement a unified, governance-first architectural framework to reliably scale Generative AI from POC experiments to fully compliant enterprise systems.
- Apply 15 production-ready architectural blueprints to unlock significant value and achieve 70-80% efficiency gains across core enterprise use cases.
- Enforce essential technical governance controls--such as DLAC, retrieval-time access checks, and immutable audit logs--to ensure complete auditability and legal defensibility.
- Design and deploy autonomous Agentic AI systems with strict safeguards, including policy-as-code enforcement, accountability layers, and robust human-in-the-loop protocols.
Who this book is for:
The book is designed for C-suite leaders--including CDOs, CTOs, CIOs, CISOs--along with Enterprise Architects and Directors of AI/ML Engineering who are responsible for scaling AI across the enterprise. It also serves Compliance Officers, Risk Managers, and senior ML engineers deploying LLMs in highly regulated sectors such as finance, pharma, manufacturing, and legal. Readers should have foundational knowledge of enterprise IT or cloud architecture and a basic understanding of Generative AI and LLMs.
商品描述(中文翻譯)
當 Generative AI 從早期的熱潮轉向實際應用，組織正面臨關鍵的十字路口。高階主管要求快速獲得投資報酬，但部署工作往往因風險、合規與治理障礙而停滯不前；此外，未受管理的 Shadow AI 系統日益增加，也讓問題更加嚴峻。《Governance-First AI》正面迎擊這項矛盾，為希望超越炒作、邁向安全且可擴充採用的領導者，提供實用且可立即執行的指南。
本書透過統一的治理優先（governance-first）方法，彌合上述落差，協助企業有信心且負責任地將 Generative AI 落實至實際營運。書中探討 MIT NANDA 一項具指標性的研究發現：95% 的企業 Generative AI 計畫未能帶來可衡量的報酬；並追溯其根本原因：解決方案架構與企業治理之間存在系統性的脫節。作者提出一套完整的架構，以實現「Certified Operational Velocity」（認證營運速度），協助組織從實驗性試點，逐步發展為符合規範且達到正式生產等級的系統。
本書的核心，是將嚴謹的治理原則——包括平台工程（platform engineering）與平台即產品（platform-as-a-product）的思維——轉化為 15 種具體的架構模式。這些藍圖涵蓋企業 AI 的各項挑戰，包括檢索增強生成（retrieval-augmented generation, RAG）部署、內容脈絡管理（context management）、舊有系統整合，以及自主 Agentic AI 系統所需的可稽核控制機制。每一章都提供逐步指引，協助建構安全、可衡量且面向未來的 AI 能力。
透過導入統一的治理優先架構，並搭配可重複使用且適合正式生產環境的應用案例，《Governance-First AI》成為解鎖 AI 真正影響力的關鍵指南。本書協助組織在不犧牲合規性的前提下，實現生產力的轉型性提升，同時為安全、自主的 AI 時代做好準備。
您將學會：
• 實作統一的治理優先架構框架，可靠地將 Generative AI 從 POC 實驗擴展至完全符合規範的企業系統。
• 應用 15 種適合正式生產環境的架構藍圖，釋放可觀價值，並在核心企業應用案例中實現 70% 至 80% 的效率提升。
• 強制執行必要的技術治理控制措施，例如 DLAC、擷取時存取檢查（retrieval-time access checks）與不可變更的稽核日誌（immutable audit logs），以確保完整的可稽核性與法律抗辯能力。
• 設計並部署具備自主能力的 Agentic AI 系統，採用嚴格的防護機制，包括政策即程式碼（policy-as-code）強制執行、責任歸屬層，以及健全的人在迴路（human-in-the-loop）協定。
本書適合對象：
本書適合負責在整個企業中擴展 AI 應用的 C-suite 領導者，包括 CDO、CTO、CIO、CISO，以及企業架構師（Enterprise Architects）和 AI/ML 工程總監（Directors of AI/ML Engineering）。本書也適合在金融、製藥、製造與法律等高度受監管產業中部署 LLM 的合規主管（Compliance Officers）、風險經理（Risk Managers）及資深 ML 工程師。讀者應具備企業 IT 或雲端架構的基礎知識，以及 Generative AI 與 LLM 的基本理解。
作者簡介
Dr M Maruf Hossain, named among the 2024 Global Top 100 Innovators in Data and Analytics, is Australia's leading Fractional Chief AI Officer (CAIO), providing strategic leadership in artificial intelligence to boards, consulting partnerships, and technology firms. He holds a PhD in Artificial Intelligence from the University of Melbourne, complemented by GAICD governance credentials.
Dr Hossain's executive experience spans the enterprise and public sectors. As Chief AI Strategist at 42 Consulting.AI, he supports organisations worldwide in building sustainable AI capabilities. Previously, as Vice President of Data Science at ANZ Bank, he led large-scale AI adoption across banking operations. His earlier roles with Credit Clear, Telstra Global, IBM Global Business Services, Infosys Consulting, and the Australian Government involved establishing Data and AI Centres of Excellence and embedding responsible AI frameworks across regulated industries.
Author of several research papers, Dr Hossain is recognised for aligning AI strategy with measurable organisational and governance outcomes.
Ahilan Ponnusamy currently works as a specialist for Application Platform at Red Hat APAC. He enjoys working with customers on Hybrid cloud architectures and cloud-native application development and delivery practices. He previously completed a Master of Computer Applications degree at Madurai Kamaraj University in India. His work history includes Philips CE in Eindhoven, Netherlands; BEA Technologies as a member of Customer Centric Engineering and support in India and the USA; Pre-sales Tech-lead for the cloud platform team at Oracle USA; Principal platform engineer at VMware; and Global Architect at Dell Technologies Singapore.
作者簡介(中文翻譯)
M Maruf Hossain 博士入選 2024 年全球資料與分析領域百大創新者，是澳洲領先的 Fractional Chief AI Officer（CAIO，分時制首席 AI 長），為董事會、顧問合作夥伴與科技公司提供人工智慧策略領導。他擁有 University of Melbourne 的人工智慧博士學位，並具備 GAICD 治理資格。
Hossain 博士的高階主管經驗涵蓋企業與公共部門。擔任 42 Consulting.AI 的 Chief AI Strategist 期間，他協助全球各地的組織建立永續的 AI 能力。此前，他曾任 ANZ Bank 的 Vice President of Data Science，負責領導銀行營運的大規模 AI 導入。他早期曾任職於 Credit Clear、Telstra Global、IBM Global Business Services、Infosys Consulting 及 Australian Government，工作內容包括建立 Data and AI Centres of Excellence，以及在受監管產業中導入負責任 AI 框架。
Hossain 博士發表過多篇研究論文，並因能將 AI 策略與可衡量的組織及治理成果相結合而廣受肯定。
Ahilan Ponnusamy 目前任職於 Red Hat APAC，擔任 Application Platform 專家。他樂於與客戶合作，處理 Hybrid cloud 架構，以及 cloud-native 應用程式的開發與交付實務。他先前於印度 Madurai Kamaraj University 取得 Master of Computer Applications 學位。他的工作經歷包括在荷蘭 Eindhoven 任職於 Philips CE；在印度與美國擔任 BEA Technologies 的 Customer Centric Engineering 及支援團隊成員；在 Oracle USA 擔任 cloud platform 團隊的 Pre-sales Tech-lead；在 VMware 擔任 Principal platform engineer；以及在新加坡 Dell Technologies 擔任 Global Architect。
