Skip to content

How to Read This Handbook

Version: 1.1 Last updated: February 2026

This handbook is comprehensive by design. You do not need to read it cover-to-cover. Use the paths below based on your role and immediate needs.

Quick Start: By Role

Product Managers

You need to understand what is possible, what is risky, and how to scope agent projects.

Essential reading (2–3 hours):

Chapter Sections What you will learn
Chapter 1 §1.1–1.2 Strategic context: why agentic AI, why now
Chapter 3 §3.1 How to select problems suited for agents
Chapter 10 §10.1–10.2 UX patterns that build user trust
Chapter 11 §11.0 Production goals and non-negotiables
Chapter 14 §14.2 Complete HR Leave example

Reference as needed:

  • Chapter 12 for governance and compliance requirements
  • Chapter 13 for team structures and roles
  • §14.6 for failure case studies (what can go wrong)

Software Engineers

You need to build agents that work reliably in production.

Essential reading (4–6 hours):

Chapter Sections What you will learn
Chapter 2 All Core concepts: agent loop, tools, memory, scaffolds
Chapter 4 All Scaffold patterns: how to structure agent behavior
Chapter 5 All Harnesses: how to test agents systematically
Chapter 9 All Best practices: prompt, tool, context patterns
Chapter 11 §11.1–11.5 Production: rollout, observability, resilience

Reference as needed:

  • Chapters 6–7 for deep evaluation methodology
  • Chapter 8 for automated optimization of prompts and trajectories
  • Appendix B for complete HR Leave reference implementation
  • §9.6 for model selection and fallback patterns

Architects

You need to design agent systems that scale and integrate with enterprise infrastructure.

Essential reading (4–5 hours):

Chapter Sections What you will learn
Chapter 2 §2.1–2.4 Architecture layers: scaffold, runtime, orchestrator
Chapter 4 §4.1–4.3 Layer separation and scaffold selection
Chapter 9 §9.5–9.6 Orchestration and model routing patterns
Chapter 11 All Production architecture and operations
Chapter 12 §12.2–12.3 Governance architecture

Reference as needed:

  • Appendix A for production challenges taxonomy
  • §12.3.6 for multi-tenant considerations
  • Chapter 13 for team structures and ownership boundaries

Governance, Risk, and Compliance

You need to understand how to evaluate and govern agent systems.

Essential reading (3–4 hours):

Chapter Sections What you will learn
Chapter 1 §1.3 Risk landscape for agentic AI
Chapter 3 §3.7–3.8 Oversight and safety principles
Chapter 6 §6.1–6.3 Evaluation foundations and metrics
Chapter 12 All Governance framework and Singapore regulations
Chapter 14 §14.6 Failure case studies

Reference as needed:

  • Chapter 10 for UX and human oversight patterns
  • Appendix A for governance checklists and templates
  • §11.7 for production readiness gates
  • Chapter 12, §12.3 for Singapore framework mappings (Model AI Governance Framework, AI Verify, MAS FEAT)

ML/AI Engineers

You need to optimize models, prompts, and evaluation within agent systems.

Essential reading (5–6 hours):

Chapter Sections What you will learn
Chapter 2 §2.5–2.7 Context engineering, memory, agent loop
Chapter 6 All Evaluation foundations: metrics, judges, datasets
Chapter 7 All Evaluation playbook: tooling and patterns
Chapter 8 All Optimization: ASI, prompt evolution, trajectory search
Chapter 9 §9.1–9.4 Prompt, tool, skill, and context patterns
Chapter 11 §11.6 Cost and latency engineering

Reference as needed:

  • Chapter 4 for scaffold design patterns
  • Chapter 5 for harness implementation
  • §9.6 for model selection patterns

Engineering Managers and Tech Leads

You need to build and lead agent teams effectively.

Essential reading (3–4 hours):

Chapter Sections What you will learn
Chapter 1 All Strategic context and landscape
Chapter 3 §3.1, §3.16–3.17 Problem selection and principles summary
Chapter 11 §11.0, §11.7 Production mindset and gates
Chapter 13 All Teams, roles, and operating models
Chapter 14 §14.6 Failure case studies

Reference as needed:

  • Chapter 12 for governance requirements
  • Appendix A for what blocks scale

Quick Start: By Task

"I need to decide if we should build an agent"

Read:

  1. Chapter 3, §3.1 — Problem selection criteria
  2. Chapter 11, §11.0.4 — Pragmatic assessment framework
  3. §14.6.1–14.6.3 — Learn from failures

Key question to answer: Is this adaptive multi-step work under uncertainty, or could a deterministic workflow handle it?

"I need to design an agent from scratch"

Read:

  1. Chapter 4 — Scaffold selection and design
  2. Chapter 9 — Best practice patterns
  3. §14.2 — HR Leave as a complete example

Deliverable: Scaffold spec with steps, tools, stopping conditions, and harness hooks.

"I need to evaluate an existing agent"

Read:

  1. Chapter 6 — Evaluation foundations (metrics, judges)
  2. Chapter 7 — Evaluation playbook (tooling, patterns)
  3. Chapter 5 — Harness design

Deliverable: Evaluation plan with metrics, datasets, judges, and thresholds.

"I need to optimize an agent"

Read:

  1. Chapter 8 — Optimization modes, ASI, algorithms
  2. Chapter 7 — Golden set management for train/val splits
  3. §8.10 — HR Leave optimization playbook

Deliverable: Optimization pipeline with seed prompt, evaluator, dataset, and validation strategy.

"I need to take an agent to production"

Read:

  1. Chapter 11 — Full production chapter
  2. Appendix A — Challenges taxonomy
  3. §11.7 — Production readiness gate

Deliverable: Production readiness checklist, rollout plan, observability setup.

"I need to set up governance for agents"

Read:

  1. Chapter 12 — Governance framework
  2. Appendix A — Checklists and templates
  3. §12.2 — Checkpoint templates

Deliverable: AI dossier template, lifecycle gates, RACI for your organization.

"I need to understand why our agent failed"

Read:

  1. §14.6 — Failure case studies
  2. §11.4 — Observability and replay
  3. §7.5 — User research and feedback loops

Approach: Map your failure to the taxonomy, identify root cause pattern, apply mitigation.

"I need to improve cost or latency"

Read:

  1. §11.6 — Cost and latency engineering
  2. §9.6 — Model selection and routing patterns
  3. §8.8 — Multi-objective Pareto optimization

Deliverable: Cost anatomy breakdown, model routing plan, caching strategy.

"I need to build a multi-region or multi-tenant agent"

Read:

  1. §14.2.7 — Internationalization (HR Leave example)
  2. §12.3.6 — Multi-tenant considerations

Deliverable: Jurisdiction mapping, tenant isolation design, evaluation stratification.

Full Handbook Reading Order

For complete coverage, this order builds understanding progressively:

+-------------------------------------------------------------+
|  FOUNDATION                                                  |
|  Ch 1: Why --> Ch 2: What --> Ch 3: Principles               |
+-------------------------------------------------------------+
                            |
                            v
+-------------------------------------------------------------+
|  DESIGN                                                      |
|  Ch 4: Scaffolds --> Ch 5: Harnesses                         |
+-------------------------------------------------------------+
                            |
                            v
+-------------------------------------------------------------+
|  MEASUREMENT & OPTIMIZATION                                  |
|  Ch 6: Eval Foundations --> Ch 7: Eval Playbook               |
|  --> Ch 8: Optimization                                      |
+-------------------------------------------------------------+
                            |
                            v
+-------------------------------------------------------------+
|  PRACTICE                                                    |
|  Ch 9: Best Practices --> Ch 10: UX Patterns                 |
+-------------------------------------------------------------+
                            |
                            v
+-------------------------------------------------------------+
|  OPERATIONS                                                  |
|  Ch 11: Production --> Ch 12: Governance                     |
+-------------------------------------------------------------+
                            |
                            v
+-------------------------------------------------------------+
|  ORGANIZATION                                                |
|  Ch 13: Teams                                                |
+-------------------------------------------------------------+
                            |
                            v
+-------------------------------------------------------------+
|  EXAMPLES                                                    |
|  Ch 14: Case Studies (including failures)                    |
+-------------------------------------------------------------+
                            |
                            v
+-------------------------------------------------------------+
|  REFERENCE                                                   |
|  Appendices: Templates, Checklists, Glossary                 |
+-------------------------------------------------------------+

Chapter Dependency Map

Use this to understand what each chapter assumes you know:

Table 1: Chapter prerequisites and downstream chapters

Chapter Assumes Enables
Ch 1 — All chapters
Ch 2 Ch 1 Ch 3, 4, 5, 9
Ch 3 Ch 2 Ch 4, 6, 10, 12
Ch 4 Ch 2, 3 Ch 5, 8, 9, 11
Ch 5 Ch 2, 4 Ch 6, 7
Ch 6 Ch 3, 5 Ch 7, 8, 11
Ch 7 Ch 6 Ch 8, 12
Ch 8 Ch 6, 7 Ch 9, 11
Ch 9 Ch 2, 4 Ch 11
Ch 10 Ch 3 Ch 11
Ch 11 Ch 4, 6–10 Ch 12, 13
Ch 12 Ch 3, 11 Ch 13
Ch 13 Ch 11, 12 —
Ch 14 All —

Appendix Quick Reference

Appendix Contents Primary Audience
A Production Challenges, Checklists and Templates Architects, Engineers, All roles
B HR Leave Agent complete reference Engineers, Architects
H Glossary All roles
E Cross-reference index All roles
F MAS AIRG mapping Governance, Risk, Compliance

End of reading guide