Product Factory Platform of Work Book Trust Login Get in touch
95% of AI pilots never reach production. The technology was never the reason.

Five Steps to
Governed AI

A discipline for firms where the cost of being wrong is higher than the cost of being slow.

The same AI is available to every firm at the same price. The only advantage left is what your firm knows, and whether you can prove that AI followed it.

85 pages 12 chapters Free to read ~45 min read
Five Steps to Governed AI - Book Cover

The cost of intelligence fell 300x.
The cost of being wrong did not.

AI is available to every firm at near-zero cost. When the technology is identical, it stops being an advantage. What remains is whether your firm can prove the AI followed its rules.

$50K-$150K
A year in fee errors, per firm. Found at the annual review, not before.
Industry estimate
95%
Of AI pilots never reach production. The model was never the problem, so the money never reaches the operations that actually cost something.
MIT Nanda, The GenAI Divide, 2025
40 hours
Of senior time, every time someone asks how you know and the answer is not already assembled.
Practitioner interviews
The Discipline
01
Decide
What to automate and what to leave alone
02
Ground
Write the rules the machine must follow
03
Calibrate
Test against your team's judgment
04
Prove
Score every run, produce evidence
05
Operate
Run in production with governance
Lower Operating Cost
Defensible Output
Capacity Without Hiring
What this book is

A discipline. Vendor-agnostic. It works with any model, any team, any integrator. We publish it freely because the industry needs a standard, and the firms that adopt it first set the terms.

Five steps, in an order that matters. Each one removes a specific cost or risk from your operations.

What this book is not

A benchmark study. We did not measure firms against each other, and we make no claim about how many have adopted any of this. Every example is constructed. No client, firm or transaction is real.

12 Chapters
01Same AI, Different Outcome

The cost of AI fell 300x. The cost of being wrong did not.

02Where the ROI Disappeared

Why 95% of AI investments stall.

03From Pilot to Operating Leverage

The five-step progression from experiment to production.

04Step One: Decide

What to automate and what to leave alone.

05Step Two: Ground

Turn unwritten judgment into enforceable policy.

06Step Three: Calibrate

Tune until the machine agrees 85-95% of the time.

07Step Four: Prove

Score every run. Produce evidence.

08Step Five: Operate

Go live with governance and evidence from day one.

09The Discipline Applied

Eight operations, end to end.

10Factory: From Judgment to a Running System

Spec to running agent in one conversation.

11Put It Into Practice

Worksheets and checklists for each step.

12Where To Go From Here

See the discipline running on real accounts.

A Note From The Founder

Most firms that buy AI get exactly what they paid for: a system that works. What they do not get is a system they can stand behind.

That gap, between "it works" and "I can prove it works," is where firms lose money, lose clients, and lose time they will never get back. Not because the AI failed. Because no one built the step where the output gets proven. We wrote this book to set out the discipline that closes that gap. It is not a product pitch. The discipline works regardless of who builds it.

Rishi Kulkarni
Co-Founder & CEO, PitCrew

Read the full discipline

85 pages. Five steps. Each one removing a specific cost or risk from your operations.