Chapter 1: Start with a Broken Moment
The Move
Start with one moment that keeps going wrong.
When I started unQbd, pronounced "uncubed," I was looking at a reading experience that had been split into pieces.
You could buy a book on Amazon. If you wanted the audiobook, you went to Audible. If you wanted conversation, you went to Goodreads. If that story became a show, you searched a streaming service or YouTube.
Each piece lived somewhere else. The pieces did not know about one another.
That was the problem I kept coming back to. Books were treated as isolated products, but readers moved between reading, listening, watching, and talking.
I did not begin with a complete platform plan. I began with a moment that felt broken.
That distinction matters. "Build a reading platform" is a solution. "A reader leaves the book to find every other part of the experience" is a problem.
Your first job is not to name an app. Your job is to describe one moment clearly enough that another person can recognize it.
Use this sentence:
When this person tries to do this, this friction keeps happening, so this cost follows.
Your result is one Broken Moment. It will become the first half of your Problem Snapshot in Chapter 2.
Why It Matters
AI can build the wrong idea very quickly.
That sounds obvious, but it changes the starting point. Building used to require enough time and skill that many weak ideas never became products. Now a vague idea can become a polished screen before anyone asks whether it should exist.
Fast building does not rescue a weak problem. It hides the weakness behind a better-looking product.
This is why the book uses the Builder Loop:
- Observe: find out what actually happens.
- Define: choose the problem and the next result.
- Direct: give the agent a bounded job.
- Prove: check the result against the brief.
- Learn: separate what the evidence shows from what you assumed.
- Decide: choose whether to continue, change, or stop.
You will use this loop at different sizes, from one interface change to a business decision. This chapter starts with Observe.
A useful broken moment has five parts:
- A trigger: something starts the moment.
- An action: the person tries to make progress.
- Friction: something slows, blocks, confuses, or repeats.
- A visible cost: time, money, effort, uncertainty, delay, or lost information follows.
- A current method: the person does something today, even if that method is awkward.
The current method is important because most problems do not sit untouched. People use notes, messages, spreadsheets, memory, paper, or another product. A workaround tells you the person cares enough to act.
It can also tell you the problem is already solved.
Imagine someone says, "Planning is annoying." That is a complaint, not yet a broken moment.
Now make it concrete:
When I return to a plan I wrote earlier, I cannot tell which version is current, so I spend ten minutes rebuilding work I already did.
The second version gives you behavior and cost. It also gives you something to observe.
Do not make the problem sound larger than it is. Small, repeated friction can support a useful product. A giant mission statement can hide the fact that no single moment has been chosen.
The early unQbd problem becomes clearer when written as behavior:
When a reader wants to continue one story in another format, the reader leaves the current experience, searches elsewhere, and starts over with another account or service.
That statement does not prove a market. It does not prove people will pay unQbd. It only names a moment worth examining.
This is the right standard for your first chapter too.
Avoid three common shortcuts.
Shortcut 1: Starting with a feature
"I want an AI dashboard" names a feature. It does not say who needs it or what fails without it.
Shortcut 2: Starting with a category
"People need a productivity app" names a market category. It does not show a repeated action or cost.
Shortcut 3: Starting with your own excitement
You are allowed to be excited. Excitement gets you moving. It does not count as evidence that someone else has the same problem.
The Broken Moment keeps those ideas honest. It is not a final verdict. It is the first claim you will try to disprove.
Three places broken moments hide
Many useful problems appear at a handoff.
One tool ends before the person's job ends. Information moves from email to a spreadsheet, from a meeting to a task list, or from a book to another format. The person has to carry the context across the gap.
The original unQbd problem is a handoff problem. The story continues, but the reader must leave one product and reconstruct the experience somewhere else.
Other problems appear through repetition.
The person copies the same answer, rebuilds the same list, checks the same status, or explains the same context every time. Repeated work does not always need software. It does deserve inspection.
A third group appears when state is lost.
The person knows what was supposed to happen but cannot tell what did happen. A note preserves an intention but not the result. A request was sent, but no one can tell who answered. A file exists, but no one knows whether it is current.
Look for these three patterns:
- Handoff: the work crosses a gap between tools, people, or formats.
- Repetition: the same useful action must be rebuilt.
- Lost state: the person cannot tell what is current, complete, approved, or true.
They are not product categories. They are places to observe.
Frequency, cost, and consequence
A problem does not need to score high on every measure.
A small problem that happens every day can matter. A rare problem with a large consequence can matter. A frequent problem with no meaningful cost may remain an annoyance.
Write all three separately:
Do not multiply guesses into a dramatic total. If you do not know the frequency, write unknown. The unknown becomes part of the evidence work.
This keeps you from turning one frustrating afternoon into a universal claim. It also keeps you from dismissing a quiet problem that steadily consumes time or trust.
A broad vision can start with a narrow break
The original unQbd vision is broader than one handoff. It connects reading, listening, watching, and conversation. That broader vision helps explain where the company may go.
It still does not change the first product question.
If I try to solve every part at once, I must decide about content rights, accounts, media, social behavior, devices, payments, and discovery. That happens before I know which broken moment matters most.
The vision should guide the direction. The Broken Moment should control the next move.
Use that distinction for your app:
You do not betray the vision by narrowing the start. You make the first claim small enough to observe, challenge, and improve.
Practice Loop
Practice Loop is a constructed teaching case used throughout this book. It is not a real unQbd product, customer study, or sales result.
The case begins with an adult hobby learner who writes practice plans in a notes app.
One note says what the person intended to practice. Days later, the note is still there, but it does not show whether the practice happened or what the result was.
The broken moment is:
When an adult hobby learner returns to an unstructured practice note, the original intention is visible but the result is missing, so the learner cannot tell what actually happened.
That looks promising because the behavior repeats and information is lost.
Then the case adds a counterexample. Another learner uses a dated paper log with a completion mark. That person can tell what happened.
The counterexample matters. It shows that "people cannot track practice" is too broad. Some people already have a working method.
The case has not earned an app yet. It has only earned a narrower question about people whose current method preserves intention but not outcome.
Do It for Your App
Start with a frustration you have seen, experienced, or can reach directly.
Do not search for the most impressive idea. Search for a moment you can describe without guessing.
Write short answers to these prompts:
Now turn the answers into the final sentence at the bottom of the same record.
Keep the sentence free of solution words when possible. Words such as app, dashboard, platform, marketplace, automation, and AI often drag you into a product too early.
Compare these examples.
Weak:
Small businesses need an AI app for customer service.
Stronger:
When a solo business owner receives the same basic question in several places, the owner rewrites the answer each time, so replies are delayed and work is repeated.
Weak:
Students need a better planner.
Stronger:
When a student receives assignments in several systems, one deadline is copied incorrectly, so the student works from an incomplete list.
The stronger versions are not automatically good businesses. They can be tested.
Next, look for a current method. The solo owner may use saved replies. The student may use a paper calendar. Those methods can become competitors, evidence, or proof that no new product is needed.
Do not hide counterevidence. If the current method works well, record that. A problem that disappears under inspection should stop the build before it consumes more time.
Finally, name the cost without exaggeration. "This ruins everything" is rarely useful. "This adds ten minutes," "this creates a missed deadline," or "this leaves the result unknown" gives you something you can check.
Before you finish, read the Broken Moment without the cost phrase. If the problem still sounds exactly the same, the cost is probably too vague.
Then check the notes behind the sentence. They should name the current method and explain where it falls short. The sentence alone does not need every detail.
The two details keep the statement grounded. They also give you a clean way to be wrong. If the cost is smaller than expected or the current method works, the idea should shrink or stop.
Prove It
Your Broken Moment passes this chapter when:
- it names one person in one situation;
- it describes behavior rather than a feature;
- the friction repeats or has a meaningful cost;
- the current method is visible; and
- at least one assumption is still labeled as an assumption.
If you cannot pass those checks, do not ask AI to build anything yet. Rewrite the moment until another person could recognize or reject it.
Your Result
Save the one-sentence Broken Moment and the seven short notes behind it.
In Chapter 2, you will choose the first person whose real behavior can test whether this moment deserves more attention.