A field manual for working differently with AI

PROGRAM YOUR AI COLLABORATOR.

Stop prompting a tool. Start building a working relationship.

Most people are learning how to get better answers from AI. This book is about building something more useful: an AI collaborator that understands how you think, challenges your assumptions, and gets better through the work.

Program Your AI Collaborator by Trent Allen and P. Kern
SYS / 01
COLLABORATOR
ARCHITECTURE
CONTEXT →
DECISION →
CORRECTION
01 / The Problem USER / COLLABORATOR

YOU ALREADY
KNOW HOW TO
USE AI.

THAT ISN'T THE PROBLEM.

The problem begins when useful conversations keep behaving like isolated conversations.

YOU'VE PROBABLY NOTICED

AI can be remarkably useful and still leave you rebuilding too much of the relationship yourself.

01

You keep explaining things it should already understand.

RESET
02

Agreement arrives faster than useful resistance.

AGREE
03

Corrections improve the moment, but rarely improve the relationship.

LOST
04

The conversations get better. The system doesn't.

FLAT
THE PATTERN

The context disappears.
The standards reset.
The relationship starts over.

02 / The Framework

Six protocols.
One relationship.

A useful AI collaborator isn't created by one perfect prompt. It develops through structure, decisions, disagreement, correction, and repetition.

01 OA

The Operating
Agreement

Define the relationship, roles, standards, boundaries, and how you work.

02 BT

Build
the Bench

Give different jobs different capabilities, perspectives, and tolerances.

03 DG

Disagreement

Make challenge part of the relationship, not an occasional exception.

04 CC

Context &
Continuity

Preserve what matters so the work can build instead of continually restarting.

05 CR

Correction

Turn misses, mistakes, and changed decisions into useful signal.

06 RP

Repetition

Let repeated work turn principles into instinct on both sides.

02.1 / The PRINCIPLE PYAC / PRINCIPLE 01

YOU ARE NOT INSTALLING
THE FINISHED COLLABORATOR.

You are giving the relationship somewhere better to begin.

02.2 / The Difference

Better prompts improve an answer.
Better architecture improves the relationship.

01 PROMPTING
A
?
Ask
B
···
Answer
C
↻
Repeat
Short context. Short-term memory. Starts over every time.
VS.
02 COLLABORATING
01
C
Context
→
02
D
Decision
→
03
↔
Disagreement
→
04
O
Outcome
→
05
+
Correction
→
06
∞
Memory
WHAT MATTERS RETURNS TO THE NEXT DECISION
Deep context. Shared standards. Compounding results.
03 / What You Build

Build a collaborator
worth working with.

The goal isn't to make AI sound more human. It's to make the working relationship more useful.

THE SHIFT

From an AI that waits for instructions to one that understands the standards behind the work.

01

Shared context.

Give the work continuity so important decisions, constraints, and standards don't have to be reconstructed every time.

CTX
02

Useful disagreement.

Create a collaborator willing to challenge assumptions, identify weak reasoning, and tell you when an idea doesn't hold up.

DGR
03

Better calibration.

Turn corrections and outcomes into signal so the collaborator becomes better aligned with how you actually make decisions.

CAL
04

Compounding value.

Stop treating every conversation as a fresh transaction. Let repeated work make the relationship increasingly useful.

CMP
04 / Inside the Manual Field Manual / Version 1.0
PYAC / 01 FIELD MANUAL
Program Your AI Collaborator by Trent Allen and P. Kern
THE SYSTEM

Not a collection of prompts.
A framework for building the relationship.

Not more prompts.
A better way to work.

The book moves from the architecture of the relationship into the decisions, corrections, and repeated work that make it useful.

01

Build the foundation.

Establish roles, expectations, standards, boundaries, and the working agreement.

FND
02

Shape the collaborator.

Create the capabilities, perspectives, and productive friction the work requires.

SHP
03

Work differently.

Move beyond question-and-answer exchanges into decisions made with shared context.

WRK
04

Correct the system.

Use disagreement, misses, and outcomes to improve how the relationship operates.

CAL
05

Let the work compound.

Turn repetition into accumulated context, judgment, and increasingly useful collaboration.

CMP
THE OBJECTIVE

The collaborator you begin with should not be the collaborator you still have six months later.

05 / THE CONTROL DECK COMPANION APPLICATION / INCLUDED

Build the architecture.

The book teaches you how. The Control Deck gives you somewhere to build it.

Create your controls, assemble your Operating Agreement, and generate a complete installation ready for your AI.

Program Your AI Control Deck
01 / CONTROL DECK

Choose what
you want to build.

Enter the working environment for the architecture introduced in the book.

02 / OPERATING AGREEMENT

Assemble the
controls.

Build the principles, standards, rules, preferences, and protocols that govern the relationship.

Operating Agreement Builder with multiple Control Blocks
Control Block showing type persistence authority and scope
03 / CONTROL BLOCK

Define what
matters.

Define what the instruction is, how long it persists, how strongly it governs, and where it applies.

04 / GENERATED AGREEMENT

Take the
architecture with you.

Generate the complete Operating Agreement with its shared definitions already included.

Generated Operating Agreement ready to copy
05 / GET THE BOOK FIELD MANUAL / VERSION 1.0
PROGRAM YOUR AI COLLABORATOR

Stop starting
over.

Build the context, standards, and working architecture that let an AI collaborator become more useful through the work.

EDITION Digital Field Manual
FORMAT PDF + EPUB
VERSION 1.0
PRICE $26.99
GET THE BOOK →

Immediate digital delivery.

HUMAN × AI / OPERATING SYSTEM

The goal isn't better answers.
It's better work.