01 · OBSERVE
The idea was generated, not discovered
Generated ideas share a signature: they describe a product but cannot name a person currently in pain.
Second Edition · 2026
Find the right problem.
Prove it. Build it.
How to find valuable problems, validate them with AI, and build products people pay for.
The difficult part is no longer producing software. The difficult part is finding a problem that is painful enough, recurring enough, reachable enough, and valuable enough to deserve building around.
This is a practical system for finding problems, validating them, narrowing them into something solvable, testing willingness to pay, building the thinnest useful solution, measuring what happens, and deciding whether to continue or stop.
Nine stagesObserve to scale or kill
Evidence firstEvery stage has an exit condition
AI as instrumentSixteen prompts, eight roles
Working toolkitLogs, scorecards, scripts, PRD
Digital ebook · Second Edition · 2026
The real failure point
You do not have an idea problem. You have a validation problem.
Post mortems of small products are boringly consistent. The cause of death is rarely the code. It is one of six failures, and every one of them happens before a line of code is written.
02 · SCORE
The problem was real but not expensive
Plenty of things are mildly annoying. Almost none justify a new vendor, a new login, and a recurring charge.
03 · INTERVIEW
The buyer and the sufferer were different people
The junior analyst suffers, the department head buys. You get enthusiastic demonstrations and no purchase orders.
04 · SCORE
The market was reachable in theory only
Two million of them means nothing if they gather nowhere. If you cannot name the place, it is a marketing problem in disguise.
05 · SLICE
The improvement was incremental
Nobody migrates data and retrains a team for thirty per cent better. They do it for four minutes instead of four hours.
06 · SCALE OR KILL
The builder fell in love before validating
Once you have named it and imagined the landing page, contradicting evidence starts to look like a solvable objection.
Every failure above is a research failure, not a technical one. The leverage in a first product is almost entirely in the first thirty days, before anything is built.
The 10X reframe
10X does not mean bigger. It means better.
Solve one painful, recurring problem ten times better for a group narrow enough that you can name its members individually.
One problem
Not a platform, not a suite, not a category. One workflow that currently costs someone hours, money, or risk.
One clear group
Small businesses is a census category. Independent consultants in one city who file fewer than fifty applications a year is a market.
One meaningful improvement
If you cannot state the improvement as a multiple, you do not have a product yet. You have a preference.
10X is a depth instruction, not a size instruction. Go narrow enough to be certain, and improve enough to be worth switching for.
The signature framework
The 10X Framework
Nine stages from problem discovery to a real product decision. Movement is forward only after evidence, and backward whenever evidence is missing.
Find
01
OBSERVE
Collect raw problem evidence in other people's words, without judging any of it.
Exit: fifty logged observations
02
SCORE
Rank candidates against seven fixed criteria and two hard gates.
Exit: top three survive both gates
03
INTERVIEW
Test the problem against people who live it. Ask about last month, never about next year.
Exit: six of ten describe the same pain
Decide
04
SYNTHESIZE
Turn evidence into a defensible problem statement. You collect, AI organises, you decide.
Exit: problem, cost, failed alternatives
05
SLICE
Cut the problem to the one step worth solving completely, and write down the non goals.
Exit: a one page PRD exists
Prove
06
SELL
Get a paid commitment before building. Offer to solve the slice by hand, this month, for money.
Exit: someone has paid
07
BUILD
Produce the thinnest thing that delivers the outcome. Descend the build ladder only when demand forces you.
Exit: a real user did the real job
08
MEASURE
Watch activation, return, and repeat payment. Return is the number that predicts survival.
Exit: you know it person by person
09
SCALE OR KILL
Decide on criteria you committed to in advance. A fast, honest kill is a completed experiment.
Exit: you honoured your criteria
The framework is a loop, not a line. Killing at stage nine returns you to stage one with a mapped industry, a contact list, and better instincts. That is a good outcome.
Ready to stop guessing what to build?
Get this ebookFifteen chapters in four parts
What is inside the book
Part one
Find
Where valuable problems hide, and how to test one before committing months.
- Why product ideas die before the first customer
- Eight things a validated problem can become
- The Twelve Hunting Grounds
- The Problem Signal Scorecard
- Ten conversations that tell you the truth
Part two
Decide
Using AI as a thinking instrument, and cutting a large problem down to one buildable thing.
- What AI actually changes, and the Wrapper Trap
- The AI research workflow
- The Solvable Slice
Part three
Prove
Getting paid before you build, and reading the only numbers that predict survival.
- Sell before you build
- The Thinnest Possible Build
- Measure what predicts survival
- Scale or kill
Part four
Practice
The ways this goes wrong, and a workbook that puts the framework into your calendar.
- Ten failure patterns, each mapped to a stage
- The 30 Day Sprint workbook
- The Manifesto
The toolkit
This is not just a book. It is a working system.
Eight sections of templates and worksheets sit at the back of the book. They are the part you use while the work is happening, not the part you read once.
A
Prompt Library
Sixteen prompts organised by the eight AI roles: research, clustering, critique, interview synthesis, slicing, pricing, outreach.
B
Problem Log
One row per observation: source, verbatim quote, frequency, workaround, cost, and where those people gather.
C
Blank Scorecards
Seven dimensions, scoring anchors, two hard gates, and a decision band. One card per candidate problem.
D
Interview Script
Thirteen questions about past behaviour, the commitment ask, and a same day notes template.
E
One Page PRD
Problem, user and buyer, the slice, one success metric, five non goals, the riskiest assumption, acceptance criteria.
F
Kill Criteria Checklist
Eight sentences to complete with a number and a date before you begin building, and a named person to hold them.
G
Glossary
Every term as this book uses it: Solvable Slice, Wrapper Trap, hard gate, activation, return, non goal, Build Ladder.
H
Changes From the First Edition
What was retained, what was rewritten, and what was added in this edition, stated plainly.
Get this ebook
Digital ebook · Second Edition · 2026
Chapter fourteen · the workbook
From observation to a decision in thirty days.
One to two hours a day. The workbook places the framework into a calendar: each phase has an objective, a task, a required output, the 10X question to ask yourself, and the trap that catches most people at that point.
The end of the sprint is a decision made on evidence. It is not a launch, and it is not a promise about revenue.
Week one · days 1 to 14
Find and score problems
Personal inventory, written evidence hunting, community immersion, structural hunting, then one scoring sitting with both gates applied.
Week two · days 16 to 19
Research, attack, reach out
Map the landscape, run the adversarial prompt against your own conclusion, then send fifteen personalised approaches to named individuals.
Week three · days 20 to 28
Interview and synthesize
Ten conversations about what people actually did last month, notes written the same day, then one page of clustered evidence with the cost attached.
Week four · days 29 and 30
Slice, sell, decide
Cut to one buildable step, write the one page PRD, set dated kill criteria, and either make the paid offer to five people or drop the problem cleanly.
The rule of this sprint
You may not begin a phase until the previous phase's output physically exists. Not mostly done: exists, in a document you can open. That single rule is what separates a sprint from a month of intentions.
Chapter six
AI can help you build. It cannot tell you what people will buy.
AI compressed two expensive stages of product work: research synthesis and code production. Reading four hundred customer reviews used to take a week. Building a functional prototype used to require a technical co founder.
The book treats it as eight distinct collaborators, with a prompt for each, and one rule that never bends: AI organises evidence you collected. It never supplies evidence, names your market, or issues a verdict.
The Wrapper Trap is the expensive version of getting this wrong. A prompt and a text box has zero switching cost, no moat, and a supplier who is also your competitor. Defensibility comes from workflow depth, proprietary context, integration, accumulated state, and trust inside a narrow community.
What AI did not change
- DistributionNobody arrives because your product exists.
- TrustA stranger's tool touching real data is a risk decision.
- DemandEffortless building does not create wanting.
- Pricing powerIt comes from the cost of the problem.
- RetentionPeople return for outcomes, not novelty.
- DefensibilityIf it took you a weekend, it takes a competitor a weekend.
AI is an ingredient, not a business.
This book is for you if
- ✓You watch broken workflows all day and suspect some of them are businesses.
- ✓You have ideas but no reliable way to tell which problems are worth pursuing.
- ✓You want to validate before committing months of your life to building.
- ✓You are an operator, coordinator, analyst, consultant, agency professional, teacher, administrator, or a product manager in transition.
- ✓You have real exposure to how work gets done, and no shipped product yet.
- ✓You want to use AI without letting AI decide what to build.
Probably not for you if
- ✕You already run a profitable software company and want advanced scaling tactics.
- ✕You want a list of startup ideas to choose from rather than a method for finding your own.
- ✕You want revenue screenshots, case studies, and market statistics. This book deliberately contains none.
The book says this in its own opening pages. It is a method book, not a hype book, and it does not try to be suitable for everyone.
Actual pages from the book
Take a look inside
Seven pages from the second edition. Select any page to enlarge it.
Questions
Before you buy
Who is this book for?+
People with real exposure to how work actually gets done, and no shipped product yet. Operators, coordinators, analysts, consultants, agency staff, teachers, administrators, and product managers in transition who watch broken workflows all day and suspect some of them are businesses.
Is this only for software products?+
No. Chapter two sets out eight things a validated problem can become: micro SaaS, an AI product, an automation product, an internal tool, a digital product, a productized service, consulting, or a small software business. The stages from Observe to Synthesize are identical regardless of which form you eventually sell, and four questions help you choose between them from evidence.
Do I need to know how to code?+
Not for the work this book is mostly about. Six of the nine stages happen before anything is built, and the Build chapter starts at the top of a ladder where you deliver the outcome manually. When you do descend to software, the book covers stack choice and four habits for building with AI without producing code you cannot debug.
How does AI fit into the method?+
As an instrument, in eight named roles: researcher, synthesizer, critic, interview analyst, workflow analyst, prototype assistant, coding assistant, and testing partner. Three rules hold throughout: never let it invent evidence, never let it name your market, and never accept a validation verdict from it. The toolkit contains sixteen prompts built on that principle.
What is the 10X Framework?+
Nine stages: Observe, Score, Interview, Synthesize, Slice, Sell, Build, Measure, Scale or Kill. Every chapter sits on one of them, and every stage has an exit condition made of evidence rather than enthusiasm. Its purpose is to let you answer one question at any moment: which stage am I actually on?
What is the 30 Day Sprint?+
A workbook that places the framework into a calendar at one to two hours a day. Each phase has an objective, a task, a required output, a 10X question, and the trap that catches most people there. It ends on day thirty with a decision made on evidence: either a paid offer made to five people, or a written post mortem and a fresh log.
What do I receive after purchase?+
The complete digital ebook, second edition, 2026: fifteen chapters in four parts, seven diagrams, and the eight part toolkit. After payment you land on a download page and the file is yours to keep. If anything goes wrong, write to pranab.kumar.pm@gmail.com.
Is this the first edition?+
No. This is the second edition, 2026. The first edition was a method guide of roughly nine thousand five hundred words in ten chapters. This edition keeps its spine and rebuilds everything around the named framework, and the final toolkit section lists exactly what was retained, rewritten, and added.
Does the book guarantee business success?+
No, and it does not pretend to. It is a method book with no invented statistics, no fabricated case studies, and no revenue screenshots. Examples are labelled as constructed. What it offers is a disciplined way to find problems, test them, and stop early when the evidence says stop.
Ready to find the problem before you build the product?
The 10X Problem Solver
Second Edition · 2026 · Digital ebook
- Fifteen chapters in four parts
- The 10X Framework, nine stages with exit conditions
- The eight part toolkit, including sixteen prompts
- The 30 Day Sprint workbook
₹99
Get this ebookPayment is handled by an external payment provider. After a successful payment you are taken straight to the download page. Questions before you buy: pranab.kumar.pm@gmail.com
The 10X Problem Solver · Second Edition
Get this ebook{{ lightboxLabel }} · select anywhere to close