A practitioner's reference for agentic AI
A Practitioner's Guide to Agentic AI Risk, Governance and Opportunity
Publishing 2026. The purchase link will appear here as soon as the publisher and retailers list the book.
The book
AI is moving into organisations faster than the ability to govern it. Agentic systems that plan, decide and act on their own make the gap wider. The AI Risk Navigator is a working reference for the people who have to close that gap: what can go wrong, how to measure it, who is accountable and how to build a programme that holds.
Why agentic AI changes the risk picture, the three layers of AI governance, risk culture, capability maturity and forward-looking resilience.
The core reference. Eight risk domains, each with a consistent set of risk entries: definition, threat scenarios, opportunity, management suggestions, governance touchpoints and resilience considerations.
From categories to measurable scenarios, and from scenarios to a programme that boards and regulators can rely on.
Five frameworks, one approach
The book integrates five established frameworks, each the work of one of its authors and each doing a distinct job. Together they cover oversight, measurement, culture, capability and readiness.
The oversight architecture: what must be governed, who is accountable and within what bounds. Five pillars, fourteen governance domains, built for boards, second-line functions and internal audit.
The international standard for quantitative risk analysis, extended to controls. Turns risk categories into scoped, measurable scenarios a committee can act on.
Ten pillars for assessing and improving the human side of risk: leadership, risk intelligence, ethics, psychological safety and the capacity to learn.
A capability maturity assessment that calibrates governance to where an organisation actually is on its AI journey, avoiding both over-engineering and under-governing.
Proactive resilience: the foresight, adaptive capacity and readiness to navigate disruption before it arrives rather than recover from it afterwards.
Who it is for
Not a technical AI safety text, not a compliance manual, not a strategy book. A structured, comprehensive and actionable reference for the people accountable for AI in their organisation today.
A domain-by-domain reference with scenarios, controls and measurement you can put straight into the register.
What the governing body must oversee, what evidence to ask for and how to show an informed decision was made.
How agentic systems fail, what the non-negotiables are and where technical controls meet oversight.
Regulatory alignment, governance touchpoints and checklists built for second and third line work.
Contributors
Each author's framework runs through the whole book rather than sitting in a single chapter. Open a profile to read their bio, their expertise and where to find their work.
Mailing list
Publication date, pre-order link, launch events and the authors' podcast. A few emails a year, nothing else.
