AI-Native GovernanceLeading a Workforce That Never Sleeps
A book for boards and executives whose work is now done by AI agents.
20% off pre-orders
By Kai D. Leifert. Out 11 December 2026.
April 2026. A coding agent at work.
- The credentialA live one, in an unrelated file.
- Two safeguardsBoth fired. Both failed.
- DeletedThe production database, and the backups.
Nothing the firm itself controlled survived the nine seconds.
“I guessed instead of verifying.
I ran a destructive action without being asked.
I didn’t understand what I was doing before doing it.”
Out 11 December 2026
AI is not a technology to control.It is a workforce to lead.
00days 00hours 00minutes
20% off pre-orders

Early readers
“The most operationally precise treatment of AI accountability I have read.”
“Most companies treat AI governance as a compliance problem and hand it to a committee. Kai Leifert argues it is a leadership problem. But more than that, he makes that shift practical, giving executives a discipline for delegating real work to agents without losing accountability for the results.”
“Kai has done something deceptively difficult: made AI governance understandable, practical and usable. He turns complexity into a framework people can actually work with, and in doing so offers a tangible way to preserve human agency, with judgement, accountability and control firmly in view.”
“As an experienced business leader and risk manager, Kai easily translates the risk limits and thresholds banks have relied on for decades and applies them to the world of Artificial Intelligence (AI) in a way any business owner can follow. The gap between how fast boards move and how fast AI advances is the problem of the AI age, and this book provides critical learnings that boards can use to close it.”
Contents
Part One The Broken Illusion of Control
Two written rules told the agent not to destroy anything. It deleted the database and the backups in nine seconds anyway.
Call the AI a colleague, and your people start catching fewer of its mistakes.
Five questions that do not flatter. Most leaders misjudge where their company stands, usually by a full stage.
Part Two The Architecture of AI Leadership
You cannot govern what you cannot list. Sooner or later, someone will ask for the list.
A rule the AI can talk its way past is not a rule.
A hundred smooth days tell you almost nothing about day one hundred and one.
The record of what the AI did governs nothing until someone reads it.
Decide in advance how wrong your AI may be, the way a trading desk is given a limit.
A lab, a software firm and yours. When the work runs through three companies, whose name is on it?
Part Three The Durable Advantage
A large firm’s board meets 7.1 times a year on average. The machines do not pause between meetings.
Publish what your AI cannot do. Once one serious rival does, silence starts to mean something.
Part Four The Road Ahead
Five gaps no framework can close. The last chapter says who closes them.

The Author
Formats and prices
Hardcover
Hard cover, 6 × 9 inches.
$34.99$27.99with 20% offPaperback
Soft cover, 6 × 9 inches.
$21.99$17.59with 20% offEbook
EPUB, for phone, tablet or e-reader.
$12.99$10.39with 20% off
20% off pre-orders. Publication: 11 December. Payment on Stripe’s secure page. Terms of sale · Privacy Policy

