01
Intent Spec
What the job is. Who's doing it, under what pressure, and what done means in their language. Not “improve budgeting UX,” but the real job with the real stakes.
● INTENT · A BOOK BY JEFF NICHOLSON
You shipped it.It works.It still missed.
How to Build Products That Last in the AI Era
If you've ever shipped something on time that didn't matter, the question is why that keeps happening, and what to replace it with. That's the book.
001 / INDEXSCROLL ↓
002
The Argument
Every product starts with a real person trying to get real work done. By the time it ships, that person is gone, replaced by a mascot stitched together from assumptions, handoffs, and rituals that reward translation over truth. Intent shows how to close that gap, using AI not to speed up the distortion but to preserve the signal.
01
Context dies in translation
Every handoff is compression. The user tells support, support tells a PM, the PM writes a roadmap, and the parts that mattered most, the ones hardest to summarize, fall out at every pass. Nobody lies. The user just quietly leaves the room.
02
Keep receipts
A summary can be reinterpreted into something safe. A receipt can't: a clip of the user failing the task, a quote that stings, a number that moved. Proof of a real moment travels with the work and keeps it honest.
03
Memory compounds
Most orgs relearn the same lesson every 18 months, because decisions dissolve into Slack threads and departed employees. Memory built as infrastructure is the one asset nobody can sell back to you. AI doesn't replace judgment. It replaces amnesia.
003
Intent
Intent
How to Build Products
That Last in the AI Era
Jeff Nicholson
BUSINESS · PRODUCT · DESIGN
Intent shows how to close that gap: how to build product systems that keep user context alive from the first complaint to the live screen, using AI not to speed up the distortion but to preserve the signal. It is for the builders who feel that gap and want to close it.
17
CHAPTERS
3
OBJECTS
1
ACCOUNTANT
PaperbackeBookAudiobook
The argument, three lengths
ONE ARGUMENT · THREE LENGTHSThe same argument at eighteen words, sixty-six, and a hundred and forty-seven. Read the one you have time for.
42,772 WORDS → 66
User intent gets lost in translation. It's not an effort problem, it's structural: every handoff is compression, and compression drops the parts that matter most, because they're the hardest to summarize. The whole model reduces to three objects: what the job is, what's true, what happened. AI's real value isn't faster paperwork. It replaces amnesia. The AI is the infrastructure. The builder is still the hero.
004
Inside the book
It's 9:47 on a Thursday night and Sue, an accountant, is still at her kitchen table fighting a budgeting tool. Her grandson had a game tonight. She missed it. By the time her frustration becomes “improve the budgeting experience” on a roadmap, she is gone. Sue is how the book stays honest: every framework has to survive one real person trying to close the books.
THE SPINE
PLUS AN INTRO, AN INTERLUDE, AND TWO APPENDICES- 01Context Preservation Systems
- 02The Polite Distortions Catalog
- 03The Reality Loop
- 04The Handoff Tax
- 05Evidence Ladders
- 06Intent Over Output
- 07Builders, Not Roles
- 08An AI-Native SDLC
- 09The System: Orchestration and Metrics
- 10How This Fails
- 11The Org: Memory
- 12The Org: Power
- 13Not Agile, But Faster
- 14How to Start
- 15From Pilot to Operating System
- 16A Living System
- 17The Work You Don't Do Yet
THREE OBJECTS
THE JOB · THE TRUTH · THE LESSON02
Receipt
What's true. A clip of a user failing the task, a quote that stings, a number that moved. Proof of a real moment, so reality can't be sanded into something safe.
03
Memory Card
What happened. What you believed, what you tried, what it did, and what you'd tell the next team, written so someone finds it before repeating the mistake.
The templates from the book, plus three Claude skills that run them. Free.
GET THE TOOLKIT →005
Field Notes
Each one takes a single idea from the book and makes the case in about 800 words. Start anywhere.
006
The Author
Jeff Nicholson is a product strategist and builder. He has led product and design teams and built AI-native systems for turning real user intent into working software. He writes and consults at the intersection of product, AI, and organizational design.
Developer
UX Designer
Design Manager
Director
SVP
Product & design
Author
Intent, 2026
007
Contact
If your team ships on time and still misses, that's the gap I help close. Intent audit or a six-week pilot on one real job.