A new book by Remko Deenik

Recompose the Work

Why Better Organization Design Beats Better AI Models

AI makes tasks faster. It does not fix the work between them.

The central argument

The biggest delays are no longer inside tasks. They live between them.

Faster tools improve individual output. They do not resolve slow handoffs, unclear authority, brittle policies, or judgment that remains unwritten.

The advantage belongs to the organization that can redesign how it works faster than the market changes what it demands.

A method for adaptability

Start with the work, not the tool.

01 / CAPABILITIES

Define what the organization must be able to do.

Begin with outcomes and enduring capabilities before choosing systems, roles, or models.

02 / CONTRACTS

Make functions explicit and composable.

Clarify authority, policy, information, inputs, outputs, and the exceptions that teach the system.

03 / AUTONOMY API

Put judgment where the work happens.

Hand work to people and AI through clear boundaries that make autonomy safer and learning reusable.

Ideas in motion

Current thinking

Book cover of Recompose the Work by Remko Deenik

The book

Recompose the Work

Why Better Organization Design Beats Better AI Models

AI adoption is becoming table stakes. This book offers a practical method, built around the Autonomy API, for making organizations explicit, adaptable, and ready to put intelligence to work.

For senior leaders at established companies who want AI to change how the business works, not only how fast individuals work.

Recompose the Work publishes October 16, 2026.

Ebook $12.99. Paperback $18.99.

Order options

Speaking

I work with leadership teams on what has to change in the organization before AI changes anything in the results.

  • Why better models do not fix broken operating systems
  • Designing authority for human and machine judgment
  • What leaders must make explicit before scaling AI
Portrait of Remko Deenik

About the author

Remko Deenik

I’ve spent more than twenty years working the seams: how work moves between teams, who gets to decide, and what breaks when the process doesn’t fit the situation.

I learned it helping build a Dutch managed services company from employee one to roughly a hundred people, then leading field and go-to-market teams across Europe and the U.S. I build AI applications with my teams, and the hard part is not the model or tool. It’s understanding what the user needs and designing the process around it, before any code gets written.

I wrote this book because I kept watching capable teams buy better tools and see the work between them stay the same.

Get in touch

Contact

For speaking, press, rights, or other inquiries, use the form below.