Skip to contentThe Observability LayerSearch

Standalone enterprise handbook · Edition 2.0

Agentic Responsible AI. From understanding to action.

Govern AI agents with clear permissions, observable outcomes, and safeguards you can test. A complete introduction, operating guide, and practical reference for the people building, buying, overseeing, and auditing them.

Choose your starting point

A reading path for the work ahead.

Read from the beginning or follow the decision you need to make. The glossary, examples, and source explanations are included; no earlier edition is required.

Understand

Start with the system

Learn what an AI agent can do, whose interests it affects, and where a failure can occur.

Chapters 1–4 · Definitions, scope, governance, threats

Read the foundations →
Govern

Set boundaries that hold

Define authority, protect data and memory, and give affected people meaningful review and correction.

Chapters 5–8 · Architecture, permissions, data, fairness

Explore the control design →
Implement

Test the claim against reality

Build evaluations, collect outcome evidence, plan recovery, and adapt the specifications to your workflow.

Chapters 9–18 · Evaluation, operation, applications, specifications

Start the 90-day plan →

The complete handbook.

18 chapters, 10 appendices, 14 explanatory figures, and 47 annotated source items. Each chapter opens as a readable page, with its citations and qualifications intact.

Explore the 10 reference appendices

Take it into your work

Read, adapt, and test.

45 proposed practices, 12 detailed implementation specifications, and 36 proposed acceptance tests connect the guidance to mechanisms, accountable owners, and observable evidence.

PDF · 116 pages · 2.9 MB

The reading edition

The complete book, including diagrams, glossary, implementation specifications, and annotated bibliography.

Download the PDF ↓
Markdown · 209 KB

The editable manuscript

Use the text in your reading and working notes. The diagrams are included in the separate implementation kit.

Download the Markdown ↓
ZIP · 61 KB · 10 blank templates

The implementation kit

Ten blank JSON templates support the 12 implementation specifications. The kit also includes practice and test registers, source and metric ledgers, SVG diagrams, and clearly marked synthetic examples.

Download the kit ↓How to use the templates →

Know what the evidence supports.

The source notes distinguish research preprints, official legal texts, technical documentation, and voluntary frameworks. Benchmark findings retain their study conditions and denominators.

The practices and tests are proposed designs. The examples are synthetic. Neither their inclusion nor a successful download demonstrates effectiveness in your enterprise.

Inspect the annotated sources →

A versioned reference. A growing evidence trail.

The handbook’s evidence cutoff remains 2 October 2026. Monthly reference updates collect newer published findings by control area. These supplementary updates do not revise the handbook text.

The July 2026 compendium and its 112 controls remain available. The handbook’s 45 practices use separate identifiers and an optional crosswalk.

Follow the daily news & insights →