Tarja Stephens seated on a white chair in a white shirt.

The Human-Agent Orchestrator

The management playbook for a world where AI is becoming part of everyday work.

We are the last generation to manage only humans

AI is giving people new management responsibilities faster than organizations are preparing them for.

As AI agents enter everyday workflows, leaders face questions most organizations have never had to answer before. What should AI be allowed to do? What decisions stay with people? Who reviews the work? When should someone intervene or stop the system?

The Human-Agent Orchestrator gives leaders a practical framework for managing this new reality.

The hardcover of The Human-Agent Orchestrator beside a tablet open on the foreword by Marshall Goldsmith and a phone playing the audiobook.
Marshall Goldsmith

The era of the user is ending. The era of the Orchestrator has begun.

Marshall GoldsmithFrom the foreword. Two-time Thinkers50 no. 1 leadership thinker
The Human-Agent Orchestrator held up against a car window at a harbour.
Tarja and a reader with the book in front of a stone wall.
Five women in Helsinki with a copy of the book.

About the book

We are the last generation to manage only humans. The next will oversee AI agents doing work at scale. Most leaders are not ready.

The rare organizations achieving 200% productivity gains from AI are not using better technology than those achieving 20%. They are managing it differently.

The authors learned this from failure. Across many AI deployments the pattern was the same: capable agents failed not from weak technology, but from a missing management layer.

Human-Agent Orchestration is that layer. A new discipline for designing, directing, and being responsible for the work AI agents do, built from four years of research across 432 organizations and hard lessons learned in the field.

Old leadership controlled the work. New leadership designs the system that produces it.

Three women holding The Human-Agent Orchestrator at a leadership event.
Tarja at a table with a stack of the book.
Tarja in a fireside conversation on bar stools under the slide Fast accelerating world.

Written by nine co-authors across three continents

Executives, professors, founders, and researchers, each bringing a different perspective to the same management challenge. With a foreword by Marshall Goldsmith and an afterword by Cassie Kozyrkov.

0
Co-authors
0
Years of research
0
Organizations

The book is traveling

From leadership sessions and industry events to readers around the world, The Human-Agent Orchestrator is opening conversations about how we prepare people to manage and oversee increasingly capable AI systems.

Tarja and a reader with the book in a wooden sauna lounge. Copies of the book on a table in front of an AI agents presentation. Tarja and a fellow author at an evening event. A stack of the book beside a bonsai tree. Two readers with the book on a Miami terrace. Copies of the book with gift bags at an event. A group of women leaders at a book gathering.
Tarja at a table with a stack of the book. The book displayed on a table in front of a screen. Tarja seated in a garden holding the book. A reader holding the book in front of a wall of books. Tarja speaking with a microphone beside a screen with her name. Guests at the Lamborghini Miami book evening. Tarja and a young reader holding the book.
Tarja in a garden with a copy of the book. Tarja at a table with a stack of books. Tarja on stage in a fireside conversation titled Fast accelerating world. Tarja on the big screen of the Visionary Stage in Luxembourg. Tarja and a friend with the book at the Lamborghini Miami evening. A stage screen reading We are the last generation to manage only humans.

Behind the book, a Miami story

Some of the most meaningful things begin when people with the same passion meet in person. The Human-Agent Orchestrator began with a meeting with Pascal Bornet in Miami, and grew into countless conversations and ideas.

Many of those conversations happened at Miami AI Club gatherings led by Nima Schei, which opened the door to deeper discussions with local leaders and innovators on what AI agents actually do inside a company.

From local meetings in Miami to global conversations with co-authors on three continents, the book became a joint answer to one question that feels more urgent every day: what happens to leadership when work is no longer done only by humans?

Most organizations are focused on building AI agents. Far fewer are asking how to manage them, govern them, assign work to them, measure their performance, decide their autonomy, and redesign human roles around them. That is the missing layer this book was written to address.

Tarja Stephens, Pascal Bornet and Nima Schei on a sofa in Miami, each holding a copy of the book.

Who this book is for

The Human-Agent Orchestrator is for anyone whose role is changing as AI becomes part of everyday work.

Across functions and levels, people are being given new responsibility for work that increasingly depends on AI systems, often before organizations have clearly defined how that responsibility should be managed.

Tarja signing a copy of the book for a reader at Lamborghini Miami.
Tarja Stephens and a young reader holding the book.
Tarja on the Nexus Luxembourg stage beside the line We are the last generation to manage only humans.

What readers and leaders say

4.9 out of 5 on Amazonfrom 139 reviews

A field guide, not a hype book

“The Orchestration Design Canvas alone is worth the price. It's the rare framework that's specific enough to use Monday morning and flexible enough to survive a messy real-world workflow.”

Jordan · United States
The management shift, named

“It is not really a book about AI. It is a book about management in the age of AI.”

Adrian · United States
Written by leaders in the room

“The author lineup is phenomenal again… many unique, practical angles packed into one book.”

Amazon reader · Japan
Human first, machine second

“It takes seriously the quiet cost to the people whose expertise gets absorbed into ‘agents’.”

Thomas · United States
A book you use, not just read

“I will reread it immediately to more fully absorb its frameworks, insights, and implications.”

Scott · United States
For the leader who has to make this work

“Its about Leadership not technology, its about change, adoption and ownership. Wish I had this book 2 years ago.”

Amazon reader · United Kingdom
A playbook for the next decade

“This book can be used as a playbook for the next decade of management. It diagnoses the real bottleneck in AI adoption, not models, but management design.”

Amazon reader
Copies of The Human-Agent Orchestrator stacked on a garden table beside a bonsai tree.

Key ideas in the book

  • The 70/30 RuleWhich management skills transfer to agents, and which will quietly sabotage your team.
  • The Supervision TrapWhy the way you currently work makes you the bottleneck instead of the leader.
  • The Orchestration Design CanvasSix layers to build a system that runs without you: Source, Success, Safety, Steering, Switch, and Sharpen.
  • The Autonomy DialWhen to give agents the keys, and when to keep humans at the wheel.
  • Default GovernanceAgents do not take control. Humans abdicate it.
  • The Human Layer IndexHow to detect human erosion before your best people leave.
  • The Identity ShiftHow to save your team from Scope Collapse, Mastery Vacuum, and Purpose Drift.

About the co-author

Tarja Stephens is the co-author of The Human-Agent Orchestrator, the executive advisor behind HAIMO, the Human-AI Management Office, and the founder of Career Ventures and Leaders of the Future.

A Finnish-American leader based in Miami, she works with the executives now responsible for the work AI does, and speaks on the leadership shift at summits from Davos to Miami.

Tarja Stephens speaking into a microphone on a panel in Miami.

Bring the book into your organization

Bring The Human-Agent Orchestrator into your leadership conversation.

Bulk orders and speaking engagements are available for executive teams, leadership programs, and company events.

From inside the book

The cartoons that run through the chapters, drawn for the book by Bornet et al.

Cartoon: a manager reads a promotion email listing 47 agents and 2 humans as direct reports. Cartoon: an agent fixed the same bug 200 times and is hailed as the most productive engineer. Cartoon: the top three tasks to hand to an agent, and why two of three would be a mistake. Cartoon: the model is brilliant, the data irrefutable, and Bob still says no. Cartoon: productivity is up 4,000% and Karen has not slept since Tuesday. Cartoon: an agent told it did a great job updates its status to still doing what you told me to do.