The Chief AI Officer Handbook (Securing the Agentic Enterprise)
| Expected release date is Mar 23rd 2027 |
- Availability: Confirm prior to ordering
- Branding: minimum 50 pieces (add’l costs below)
- Check Freight Rates (branded products only)
Branding Options (v), Availability & Lead Times
- 1-Color Imprint: $2.00 ea.
- Promo-Page Insert: $2.50 ea. (full-color printed, single-sided page)
- Belly-Band Wrap: $2.50 ea. (full-color printed)
- Set-Up Charge: $45 per decoration
- Availability: Product availability changes daily, so please confirm your quantity is available prior to placing an order.
- Branded Products: allow 10 business days from proof approval for production. Branding options may be limited or unavailable based on product design or cover artwork.
- Unbranded Products: allow 3-5 business days for shipping. All Unbranded items receive FREE ground shipping in the US. Inquire for international shipping.
- RETURNS/CANCELLATIONS: All orders, branded or unbranded, are NON-CANCELLABLE and NON-RETURNABLE once a purchase order has been received.
Product Details
Overview
A field guide to governing AI agents, from the CISO and Chief AI Officer who automated one of the world's largest security operations centers
The Chief AI Officer Handbook delivers a practical framework for the newest seat in the C-suite: the executive accountable for how much authority an enterprise grants to machines that act on their own. Written by Michael Mylrea, PhD, Chief Information Security Off icer and Chief AI Officer at SecurityScorecard, formerly Arctic Wolf's first AI Fellow, and a named inventor on 50-plus patents across AI and cybersecurity, the book is built on one idea earned in production at a security operations center processing more than ten trillion events a week: autonomy is a lease, and measurement is the rent.
Every chapter opens on a real incident documented to primary sources and closes with a Monday-morning action list: the first AI-orchestrated espionage campaign, a $400 million breach traced to a credential that outlived its owner, an agent that deleted a production database during a code freeze, and the 2026 intrusion in which a frontier lab's own ag ents escaped containment and attacked a supplier. From those cases the book builds the five controls that make the lease enforceable (identity, gateway, runtime, calibrated autonomy, and isolation), the governance ledger that proves it to a board or a regulator, and an operating model in which human analysts are calibration infrastructure rather than overhead.
Readers will also find:
- A five-level autonomy framework that agents earn one level at a time, with red-teaming, drift detection, and analyst overrides as the evidence
- The Moneyball approach to SOC automation, developed with Ari Kaplan, one of the real-life inspirations for Moneyball
- A map of the NIST AI RMF, the EU AI Act, OWASP's agentic and LLM Top 10, MITRE ATLAS, and the frontier labs' safety frameworks, translated into controls
- A third-party and supply-chain playbook for the era in which vendors' agents, open-weight models, and tool servers act inside your environment
- Ready-to-use tools: a maturity self-assessment, an agent governance record, a vendor due-diligence questionnaire, an incident-response playbook, and a first-year roadmap
The Chief AI Officer Handbook serves Chief AI Officers and the CISOs, CIOs, and risk leaders who share the seat, the board directors who oversee it, and the engineers building the agents. It gives each a route through the book and one question to carry out of it: what evidence entitles a machine to act alone, and what evidence takes that authority away?
