Prompt. Context. Intent.
A prompt is how you ask. Context is what the AI model sees. Intent is why you are asking at all.
- PromptThe words. How you phrase the request.
- ContextEverything around the words: what the AI model can see.
- IntentWhat all of it is for. The part that is yours.
Same task. Three different jobs.
Tell an agent "cut the grass" and it will cut the grass. Whether that was right depends on what the grass is for.
A neat lawn
Cut it every week.
A meadow for bees
Do not cut it at all.
To sell the house
Cut, edge and plant.
Same task: done, not done, or done with more, depending on what it is for.
Where intent sits.
Your semantic layer says what revenue means. Your intent layer says what you want it to be, what must not break on the way, and proves whether your agents got it there.
The ontology sits between the two. It tells an agent what exists and what it is able to do, and the intent layer tells it which of those things to do, and why. A company needs all three, and most have built only the first two.
- Intent layerWhat you want, what must not break, and the proof.Grow revenue from large customers. Never discount more than 10 percent. Checked every week.
- OntologyWhat exists, how it connects, what can be done.Customers place orders. An order can be refunded.
- Semantic layerWhat the words mean.Revenue is bookings minus refunds.
- DataThe facts, stored.Orders, invoices, payments.
Three moves.
How to turn what you want into something people and AI can act on. Three moves, three steps each: the playbook on one page, as the book ends.
- 1
Find it
Ask why until you reach what you actually want, probe it by doing, and measure the gap between what you say and do. Taught in Chapters 2 and 10.
- 2
Write it
One page, the contract, with trade-offs that cost something. Then the swap test: does someone else decide as you would? Taught in Chapter 5, tested in Chapter 6.
- 3
Keep it
Keep it current: hand it off with authority and stop conditions, review it monthly, and change it on purpose. Taught in Chapter 11.
The intent contract.
Six sections, one page. Agents read down from it. You keep it.
- PurposeWhy it exists at all.
- OutcomesWhat will be true, by when, and why.
- Trade-offsThis over that, when two good things collide.
- BoundariesNever. May. Must ask.
- MeasuresSigns of progress, and when to stop.
- Owner and decisionsWhose it is, when it is read again, and a dated log of the calls made under it.
An intent contract for a small bakery
Purpose. A bakery the town loves, that gives me and my two staff a good life.
Outcomes. Thirty more regular customers by spring, because regulars carry us through the winter.
Trade-offs. Regulars over one-time buyers. Margin over volume. Our time off over growth.
Boundaries. Never discount more than 10 percent. Never open or work on Sundays or Mondays. Agents may draft posts and emails, and must ask before spending any money.
Measures. Customers who come back three or more times a month. Stop anything that brings crowds but no repeat visits.
Owner and decisions. Me. Read again monthly. Decisions: none yet.
The swap test.
Hand your page and a hard case to someone else. If they make your call, for your reasons, the page holds.
A bakery's page said "our time off over growth". A café offered 20 percent more revenue if someone started at 4 a.m. No rule forbids it; the owner would say no.
Given the rules alone, the big AI models asked the owner. Given the whole page, they said no, for her reasons.
Start this week.
The tasks that close the chapter on the intent contract.
What to do this week
The workbook.
Twelve pages to print and write on, one for every move in the book: the magic wand, the five whys, your intent contract, the swap test, your system prompt for life.
- The magic wand
- My five whys
- My intent contract, the top
- My intent contract, the rest
- Three trade-offs that cost something
- My intent map
- A swap test
- Intent evals
- The intent audit
- A hand-off
- My system prompt for life
- My yearly review
Read it over a weekend.
Diagrams on almost every spread, samples you can copy, and things to do at the end of every chapter.
Twelve chapters, four parts.
A prologue on the last prompt you would ever write, then four parts of three chapters: what an intent is, why we call it engineering, where this is going, and getting ready, whatever your role.
Part oneWhat an intent is
- 1Intent is older than AI
- 2Tasks are not intents
- 3Life is a series of decisions
Part twoWhy we call it engineering
- 4What engineering means
- 5The intent contract
- 6The swap test
Part threeWhere this is going
- 7From autocomplete to one word
- 8When every team has agents
- 9What could go wrong
Part fourGetting ready, whatever your role
- 10The customer asked for the wrong thing
- 11Hand it off, keep it alive
- 12When the robot does the dishes
Also: the playbook on one page, six worksheets, and the full experiment, so anyone can repeat it.
It argues with itself.
An idea worth adopting can survive its strongest objections. Chapter 9 makes the case against.
A written intent is a commitment about your future: who you are trying to become, which may be different from who you have been.
About the author
Fru Nde
Fru Nde is a technologist and enterprise seller. He has spent his career selling data and AI to some of the world's largest and most strategic customers, in the rooms where big software decisions get made.
He has walked large companies through two shifts: from the data centre to the cloud, and now to AI. In a complex enterprise deal, where a decision passes through many hands before anyone signs, the question that matters most is what the customer is actually trying to do. As AI takes over more of the doing, that question becomes everyone's. Intent Engineering is his playbook for answering it: find the intent, write it down, and keep it current.
He writes at fru.dev.
Where to learn more.
Every chapter ends with a few places to go further. Here they all are.
1Why the pyramids were built: intent is older than AI
- The Common Law, by Oliver Wendell Holmesgutenberg.org
The 1881 lectures, free, where the dog line comes from, in the first lecture. The second is on criminal law.
- Man's Search for Meaning, by Viktor Franklany library
A short book on why the why matters, written by a survivor of the camps.
- A Theory of Human Motivation, by Abraham Maslowpsychclassics.yorku.ca
The 1943 paper itself, free to read, and shorter than you expect.
2Tasks are not intents: why would anyone cut the grass?
- The Toyota Production System, by Taiichi Ohnoany library
The 1978 book by the man who made the five whys part of how Toyota works.
- Competing Against Luck, by Clayton Christensenany library
The jobs-to-be-done idea: what customers hire a product to do.
3Life is a series of decisions, and intent is how you judge them
- Thinking, Fast and Slow, by Daniel Kahnemanany library
The two systems of thinking, and why fast decisions go wrong.
- Thinking in Bets, by Annie Dukeany library
A former poker player on judging decisions by their odds and not by how they turned out.
4What engineering means, and why intent needs it
- Mission commandarmypubs.army.mil
The US Army's doctrine on commander's intent, free to read. Search for Army Doctrine Publication (ADP) 6-0.
- Software Engineering, the 1968 NATO conference reporthomepages.cs.ncl.ac.uk
The report from Garmisch, edited by Peter Naur and Brian Randell, where the discipline got its name.
- Start with Why, by Simon Sinekany library
The Golden Circle, why before how before what, written for leaders.
5The intent contract: one page someone else can act on
- AGENTS.mdagents.md
An open convention for a file of standing instructions that AI agents read before they work.
- Working Backwards, by Colin Bryar and Bill Carrany library
How Amazon writes the press release before building the product: an intent written first.
7From autocomplete to one word: where the instruction is heading
- Measure What Matters, by John Doerrwhatmatters.com
The OKR story from Intel to Google, with templates.
- The Practice of Management, by Peter Druckerany library
The 1954 book that introduced management by objectives.
- Effective context engineering for AI agentsanthropic.com/engineering
An essay on what to give an AI before it starts.
- Model Context Protocolmodelcontextprotocol.io
The open standard for connecting AI to tools and data, and a model for what an intent protocol might become.
- No Silver Bullet, by Fred Brookstechreports.cs.unc.edu
The 1986 essay that split software's difficulty into the accidental and the essential: deciding what to build.
8When every team has agents: shadow intent, culture and constitutions
- Statement on the Purpose of a Corporationbusinessroundtable.org
The 2019 statement, with the list of signers.
- Claude's constitutionanthropic.com/constitution
A published example of an intent document written for a machine.
9What could go wrong: bad goals, paperclips and the case against this book
- Ethical Issues in Advanced Artificial Intelligence, by Nick Bostromnickbostrom.com/ethics/ai
The 2003 essay with the paperclip thought experiment.
10The customer asked for the wrong thing: finding what you actually want
- SPIN Selling, by Neil Rackhamany library
The research on how the best sellers ask questions.
- Effectuation, by Saras Sarasvathyeffectuation.org
How expert entrepreneurs find their goals by acting.
11Hand it off, keep it alive: delegating without disappearing
- Turn the Ship Around!, by L. David Marquetany library
A submarine captain on handing decisions down through stated intent.
- Moffatt v. Air Canada, 2024 BCCRT 149canlii.org
The tribunal's decision on who answers for what a chatbot says.
12When the robot does the dishes: what is left for you to do
- The Structure of Scientific Revolutions, by Thomas Kuhnany library
The 1962 book on how one shared picture of the world gives way to the next.
- Marienthal, by Marie Jahoda, Paul Lazarsfeld and Hans Zeiselany library
The 1933 study of a village out of work, and what work gives besides a wage.
- Augmenting Human Intellect, by Douglas Engelbartdougengelbart.org
The 1962 report that set out to make people more capable, not to replace them. Free to read.
A living book.
This book is versioned like software. Writing with AI makes revising fast, so the book improves as its thinking does, instead of waiting years for a new edition. Check the version on your copy's copyright page, and see here what changed in each one.
Current version 1.3.0
- 1.3.0Declarative systems and the why they leave out, English as the new programming language, and the history of the ladder checked against its sources.
- 1.2.0Where the ladder of abstraction began and whether it has a top, and the bicycle for the mind.
- 1.1.1A bolder cover: ENGINEERING across the full width, and the author's name set larger.
- 1.1.0Planning stories, from houses and cities to journeys and a burger menu, naming the AI, the ultimate intent, and a new cover.
- 1.0.0First release.
Read the book online.
The whole book, online. No email, no sign-up. Or buy the paperback.