What Will You Learn?
By the end of this lesson, you will be able to:
- Understand why problem scoping is the foundation of any AI project
- Use the 4Ws Canvas (Who, What, Where, Why) to define problems clearly
- Create effective problem statements using the Problem Statement Template
- Distinguish between themes, topics, and problems in AI projects
Have you ever wanted to build an AI app that helps students study better?
Great idea but before you do anything about it, you need to think that app should help them study what? All subjects or just one? Help them remember things or understand concepts? For students who are struggling or those who want to score higher?
See the problem? Your “great idea” is actually quite vague. And if you don’t know exactly what problem you’re solving, how will you know if your AI solution actually works?
This is why Problem Scoping exists. It’s the art of turning a fuzzy idea into a crystal-clear problem statement.
Think of it like visiting a doctor. If you just say “I feel unwell,” the doctor can’t help much. But if you say “I have a headache that started yesterday, gets worse when I look at screens, and happens mostly in the evening” — now the doctor knows exactly what to investigate.
Problem scoping does the same for AI projects. Let’s learn how.
What is Problem Scoping?
Problem Scoping is the first and most critical stage of the AI Project Cycle. It answers one fundamental question:
What exactly are we trying to solve?
A well-scoped problem is:
- Clear: Everyone understands what needs to be solved
- Specific: Not too broad or too narrow
- Meaningful: Actually matters to someone
- Achievable: Can be solved with available resources and technology
Without proper problem scoping, you might:
- Collect the wrong data
- Build a solution nobody needs
- Waste time and effort on the wrong approach
- Create something that doesn’t actually solve the real problem
💡 Key Insight
Problem Scoping is like setting the GPS destination before a road trip. If your destination is wrong, no amount of driving will get you where you want to go.
The 4Ws Canvas: Your Problem Definition Tool
The 4Ws Canvas is a simple but powerful framework to break down any problem. It asks four essential questions:
| Question | What It Uncovers |
|---|---|
| WHO | Who is affected by this problem? Who are the stakeholders? |
| WHAT | What exactly is the problem? What needs to be achieved? |
| WHERE | Where does this problem occur? In what context or location? |
| WHY | Why is solving this problem important? What’s the impact? |
Let’s explore each “W” in detail.
WHO: Identifying Your Stakeholders
The “Who” question helps you understand the people involved in your problem. These people are called stakeholders.
Questions to ask:
- Who faces this problem directly?
- Who will use the AI solution?
- Who benefits when the problem is solved?
- Who might be affected (positively or negatively) by the solution?
Example: AI for School Attendance Monitoring
| Stakeholder | Their Role |
|---|---|
| Students | Those whose attendance is being tracked |
| Teachers | Need to know who’s present in class |
| Parents | Want to know if their child attended school |
| School Administration | Need attendance records for reporting |
| Government | Requires school attendance data for education statistics |
🧪 Think About It
Can you think of a stakeholder who might NOT want an AI attendance system? Why might they object?
WHAT: Defining the Problem Clearly
The “What” question focuses on the actual problem you need to solve. This requires careful thinking because problems can be stated at different levels.
Moving from Theme to Topic to Problem:
| Level | Example | Clarity |
|---|---|---|
| Theme | Education | Very broad |
| Topic | Student performance | Still broad |
| Problem | Many students don’t complete homework on time | Specific and actionable |
See how we got more specific? The AI solution becomes clearer when the problem is precise.
Questions to ask:
- What exactly is going wrong?
- What outcome do we want to achieve?
- What does success look like?
- What constraints or limitations exist?
Example: AI for Traffic Management
| Level | Description |
|---|---|
| Theme | Transportation |
| Topic | Urban traffic congestion |
| Problem | Traffic signals don’t adjust to real-time traffic flow, causing unnecessary waiting at empty intersections |
The third statement gives us something concrete to solve with AI — intelligent traffic signals that adapt to actual traffic patterns.
WHERE: Understanding the Context
The “Where” question identifies the location, context, or environment where the problem exists.
Questions to ask:
- Where does this problem happen? (Physical location)
- In what situations does it occur? (Context)
- What systems or platforms are involved?
- Are there specific conditions when it’s worse?
Example: AI for Healthcare
| Where Aspect | Details |
|---|---|
| Physical Location | Rural health centers in Tamil Nadu with limited doctor availability |
| Context | During diabetic patient screenings |
| System | Eye examination equipment |
| Conditions | Problem is worse during monsoon when roads are bad and patients can’t travel to city hospitals |
The “where” tells us our AI solution needs to work in resource-limited settings with basic equipment — a crucial design requirement.
WHY: Establishing the Importance
The “Why” question justifies why this problem matters and deserves an AI solution.
Questions to ask:
- Why does this problem need to be solved?
- What happens if we don’t solve it?
- Why use AI specifically? (vs. other approaches)
- What positive impact will the solution create?
Example: AI for Detecting Crop Diseases
| Why Aspect | Answer |
|---|---|
| Why solve it? | Farmers lose 15-25% of crop yield to diseases every year |
| What if we don’t? | Food shortage, farmer income loss, increased food prices |
| Why use AI? | Traditional methods require experts who aren’t available in remote areas |
| Positive impact | Early detection can save crops, increase farmer income, ensure food security |
A strong “why” motivates everyone involved and helps prioritize resources.
Putting It Together: Complete 4Ws Canvas Example
Let’s apply the 4Ws Canvas to a real scenario.
Scenario: Your school wants to reduce electricity wastage caused by lights and fans being left on in empty classrooms.
4Ws Canvas: Smart Classroom Energy Management
| W | Question | Answer |
|---|---|---|
| WHO | Who is affected? | Students (learning environment), teachers, school management (electricity bills), environment (carbon footprint) |
| WHAT | What is the problem? | Lights and fans remain on even when classrooms are empty, wasting electricity and money |
| WHERE | Where does it occur? | School classrooms, especially during breaks, after school hours, and on holidays |
| WHY | Why solve it? | Reduce electricity bills by 20-30%, teach students about sustainability, reduce environmental impact |
From this canvas, the problem statement becomes clear:
Problem Statement: “We need an AI system that automatically detects when classrooms are empty and turns off lights and fans to reduce electricity wastage and promote sustainable practices.”
The Problem Statement Template
Once you’ve completed the 4Ws Canvas, you can create a formal Problem Statement using this template:
We need to [WHAT – action/solution] for [WHO – stakeholders] in [WHERE – context] because [WHY – reason/impact].
Examples:
Example 1: Spam Detection
“We need to automatically identify and filter spam emails for email users across all email platforms because spam wastes time and exposes users to phishing attacks and scams.”
Example 2: Traffic Management
“We need to create intelligent traffic signals that adjust to real-time traffic flow for city commuters at major intersections because fixed-timing signals cause unnecessary delays and increase fuel consumption.”
Example 3: Diabetic Retinopathy Detection
“We need to build an AI system that screens eye images for diabetic retinopathy for diabetic patients in rural Tamil Nadu vision centers because there aren’t enough eye specialists to screen everyone, and untreated cases lead to blindness.”
Common Mistakes in Problem Scoping
Avoid these pitfalls when defining your AI problem:
| Mistake | Example | Better Approach |
|---|---|---|
| Too Vague | “Improve education” | “Help Class 9 students practice mathematics through personalized problem recommendations” |
| Too Ambitious | “Create an AI that knows everything” | “Create a Q&A system that answers science questions from NCERT textbooks” |
| Solution-First Thinking | “We need a chatbot” | “Students have questions after school hours when teachers aren’t available” (then decide if chatbot is the right solution) |
| Missing Stakeholders | Only considering students | Consider teachers, parents, school administration too |
| Unclear Success Criteria | “Make students happy” | “Increase homework completion rate from 60% to 85%” |
💡 Remember
Start with the problem, not the solution. Don’t think “I want to use AI” — think “I have this problem” and then ask if AI is the right tool.
Activity: Create Your Own 4Ws Canvas
Here’s a scenario for you to practice:
Scenario: Many students in your school waste food during lunch. The canteen throws away 20-30 kg of uneaten food daily.
Create a 4Ws Canvas to define this problem:
| W | Question | Your Answer |
|---|---|---|
| WHO | Who is affected? | |
| WHAT | What is the problem? | |
| WHERE | Where does it happen? | |
| WHY | Why solve it? |
Now write a problem statement using the template.
(Suggested answers are in the Answer Key below)
Real-World Case Study: CottonAce App for Pest Management
Let’s see how proper problem scoping led to a successful AI solution.
The 4Ws Canvas:
| W | Details |
|---|---|
| WHO | Cotton farmers in India, agricultural extension officers, pesticide companies |
| WHAT | Pink bollworm pest destroys cotton crops; farmers spray pesticides blindly without knowing if pests are present |
| WHERE | Cotton fields across India, particularly during the growing season |
| WHY | Farmers lose 20-30% of crop value; excessive pesticides harm soil and health; proper timing of treatment can save crops and reduce costs |
Problem Statement:
“We need an AI system that identifies pink bollworm presence in cotton fields through image recognition for cotton farmers in India because blind pesticide spraying wastes money and harms the environment while pests still damage crops.”
Result: The CottonAce app was developed, leading to:
- 26.5% increase in farmer profits
- 38% reduction in pesticide costs
- Better crop health and environmental outcomes
This is the power of good problem scoping!
Quick Recap
- Problem Scoping is the first and most important stage of the AI Project Cycle.
- The 4Ws Canvas helps define problems using four questions: Who, What, Where, and Why.
- WHO identifies all stakeholders affected by the problem and solution.
- WHAT defines the specific problem, moving from broad themes to specific issues.
- WHERE establishes the context, location, and conditions of the problem.
- WHY justifies the importance and impact of solving the problem.
- Use the Problem Statement Template to create clear, actionable statements.
- Always start with the problem, not the solution — avoid solution-first thinking.
- Good problem scoping prevents wasted effort and leads to better AI solutions.
Previous Lesson: The AI Project Cycle Explained: 6 Steps to Build Your First AI Project
Next Lesson: Data Acquisition in AI: How to Collect, Source and Gather Data for Machine Learning Projects
EXERCISES
A. Fill in the Blanks
- Problem Scoping is the ___________________ stage of the AI Project Cycle.
- The 4Ws stands for Who, What, Where, and ___________________.
- The people affected by a problem are called ___________________.
- When moving from a theme to a problem, we need to get more ___________________.
- The “Why” question helps establish the ___________________ of solving a problem.
- A well-scoped problem should be clear, specific, meaningful, and ___________________.
- ___________________-first thinking is a common mistake where people decide the solution before understanding the problem.
- The 4Ws Canvas is a ___________________ used to break down and define problems.
- “Education” is an example of a broad ___________________, not a specific problem.
- The Problem Statement Template includes WHO, WHAT, WHERE, and ___________________.
B. Multiple Choice Questions
1. Which stage of the AI Project Cycle comes first?
(a) Data Acquisition
(b) Problem Scoping
(c) Modelling
(d) Evaluation
2. What does the “Who” in 4Ws help identify?
(a) The solution approach
(b) The stakeholders affected
(c) The location of the problem
(d) The technology to use
3. Which of the following is the most specific?
(a) Transportation
(b) Urban traffic
(c) Traffic signals don’t adjust to real-time flow
(d) City problems
4. “Solution-first thinking” means:
(a) Thinking about the problem first
(b) Deciding on a solution before understanding the problem
(c) Finding the best solution
(d) None of the above
5. Which question does the “Where” in 4Ws answer?
(a) Why is this important?
(b) Who is affected?
(c) In what context does this occur?
(d) What needs to be achieved?
6. A good problem statement should NOT be:
(a) Clear
(b) Specific
(c) Vague
(d) Meaningful
7. In the CottonAce case study, what problem did the AI solve?
(a) Weather prediction
(b) Pest identification in cotton crops
(c) Irrigation management
(d) Crop pricing
8. Stakeholders are:
(a) Only the people who create the AI
(b) Only the people who use the AI
(c) All people affected by the problem and solution
(d) Only the investors
9. Which is a theme, not a problem?
(a) Healthcare
(b) Students miss homework deadlines
(c) Traffic lights waste time at empty intersections
(d) Food gets wasted in canteen
10. The Problem Statement Template follows the format:
(a) WHO – WHAT – WHY – WHERE
(b) WHAT – WHO – WHERE – WHY
(c) WHERE – WHO – WHAT – WHY
(d) WHY – WHAT – WHERE – WHO
C. True or False
- Problem Scoping can be skipped if you already know what AI solution you want to build. (__)
- The 4Ws Canvas helps make vague problems more specific. (__)
- A stakeholder is someone who is affected by the problem or its solution. (__)
- “We need a chatbot” is a well-scoped problem statement. (__)
- The “Why” question justifies the importance of solving the problem. (__)
- It’s okay to have multiple stakeholders for a single problem. (__)
- “Education” is a specific problem ready for an AI solution. (__)
- The CottonAce app helped reduce pesticide costs for farmers. (__)
- Problem Scoping comes after Data Acquisition in the AI Project Cycle. (__)
- A problem statement should include context about where the problem occurs. (__)
D. Define the Following (30-40 words each)
- Problem Scoping
- 4Ws Canvas
- Stakeholder
- Problem Statement
- Theme (in problem scoping context)
- Topic (in problem scoping context)
- Solution-First Thinking
E. Very Short Answer Questions (40-50 words each)
- What is Problem Scoping and why is it the first step in AI Project Cycle?
- List the four questions in the 4Ws Canvas and what each uncovers.
- What is the difference between a theme, topic, and problem? Give an example.
- Why is identifying stakeholders important in Problem Scoping?
- What are three characteristics of a well-scoped problem?
- What is “solution-first thinking” and why should it be avoided?
- How does the “Where” question help in defining a problem?
- Write the Problem Statement Template format.
- Give one example of a vague problem and make it more specific.
- Why is the “Why” question important in the 4Ws Canvas?
F. Long Answer Questions (75-100 words each)
- Explain the 4Ws Canvas in detail with an example of your choice.
- Why is Problem Scoping considered the most important stage of the AI Project Cycle? What happens if it’s done poorly?
- Create a complete 4Ws Canvas for an AI system that helps students manage their study time better.
- Explain the difference between theme, topic, and problem with examples from the education domain.
- Describe the CottonAce case study and explain how proper problem scoping contributed to its success.
- What are common mistakes in Problem Scoping? How can they be avoided?
- Create a 4Ws Canvas and Problem Statement for an AI system that reduces food wastage in school canteens.
ANSWER KEY
A. Fill in the Blanks – Answers
- first — Problem Scoping is the foundation stage before any other work begins.
- Why — The 4Ws are Who, What, Where, and Why.
- stakeholders — Stakeholders are all people affected by the problem or solution.
- specific — We narrow down from broad themes to specific, actionable problems.
- importance — The Why question establishes why the problem matters.
- achievable — A well-scoped problem must be solvable with available resources.
- Solution — Solution-first thinking puts the cart before the horse.
- framework/tool — The 4Ws Canvas is a structured framework for problem definition.
- theme — Education is a broad theme, not a specific problem.
- WHY — The template includes all four Ws: WHO, WHAT, WHERE, and WHY.
B. Multiple Choice Questions – Answers
- (b) Problem Scoping — It’s the first stage that defines what we’re solving.
- (b) The stakeholders affected — Who identifies all people involved in the problem.
- (c) Traffic signals don’t adjust to real-time flow — This is the most specific and actionable.
- (b) Deciding on a solution before understanding the problem — This leads to building wrong solutions.
- (c) In what context does this occur? — Where establishes location and context.
- (c) Vague — Vague problems can’t be solved effectively.
- (b) Pest identification in cotton crops — The app detected pink bollworm in cotton.
- (c) All people affected by the problem and solution — Stakeholders include users, beneficiaries, and anyone impacted.
- (a) Healthcare — Healthcare is a broad theme, not a specific problem.
- (b) WHAT – WHO – WHERE – WHY — The template: “We need to [WHAT] for [WHO] in [WHERE] because [WHY].”
C. True or False – Answers
- False — Problem Scoping should never be skipped; it’s foundational.
- True — The 4Ws Canvas transforms vague ideas into clear problems.
- True — Stakeholders are all people affected by the problem or its solution.
- False — “We need a chatbot” is solution-first thinking, not a problem statement.
- True — The Why question explains the importance and impact.
- True — Most real problems have multiple stakeholders.
- False — Education is a broad theme, not a specific problem.
- True — CottonAce reduced pesticide costs by 38%.
- False — Problem Scoping comes BEFORE Data Acquisition, as the first stage.
- True — The “Where” context is part of a complete problem statement.
D. Definitions – Answers
1. Problem Scoping: The first stage of the AI Project Cycle where we clearly define what problem needs to be solved using tools like the 4Ws Canvas to move from vague ideas to specific, actionable problem statements.
2. 4Ws Canvas: A framework for defining problems by answering four questions — Who is affected, What is the problem, Where it occurs, and Why solving it matters. It creates clarity before building solutions.
3. Stakeholder: Any person or group who is affected by a problem or its solution, including direct users, beneficiaries, and anyone impacted positively or negatively by the AI system.
4. Problem Statement: A clear, concise description of the problem that specifies what needs to be solved, for whom, in what context, and why it matters, guiding all further development work.
5. Theme (in problem scoping): A broad area or domain of interest like “education” or “healthcare” that needs to be narrowed down to specific topics and problems before an AI solution can be designed.
6. Topic (in problem scoping): A specific area within a theme, like “student performance” within education, that is more focused but still needs to be narrowed to a concrete, actionable problem.
7. Solution-First Thinking: A common mistake where people decide on a solution (like “we need a chatbot”) before properly understanding the problem, often leading to building solutions that don’t address real needs.
E. Very Short Answer Questions – Answers
1. What is Problem Scoping and why is it the first step?
Problem Scoping defines exactly what problem we’re trying to solve. It’s first because everything else depends on it — the data we collect, the model we build, and how we evaluate success all flow from the problem definition.
2. List the four questions in the 4Ws Canvas:
WHO identifies stakeholders affected by the problem. WHAT defines the specific problem to solve. WHERE establishes the context and location. WHY justifies why solving this problem matters and its impact.
3. Difference between theme, topic, and problem:
A theme is broad (e.g., “Transportation”), a topic is more focused (e.g., “Traffic congestion”), and a problem is specific and actionable (e.g., “Traffic signals don’t adapt to real-time traffic flow”). We move from general to specific.
4. Why is identifying stakeholders important?
Identifying stakeholders ensures we understand everyone affected by the problem and solution. It helps design solutions that work for all users, avoid unintended negative impacts, and gather diverse perspectives for better outcomes.
5. Three characteristics of a well-scoped problem:
A well-scoped problem is specific (not too broad or narrow), meaningful (matters to stakeholders), and achievable (can be solved with available technology and resources).
6. What is solution-first thinking?
Solution-first thinking means deciding on a solution (like “we need an app”) before understanding the problem. It should be avoided because it often leads to building solutions that don’t address the real problem.
7. How does “Where” help in defining a problem?
The “Where” question establishes the context, location, and conditions of the problem. It reveals environmental constraints, specific situations when the problem occurs, and requirements the solution must meet.
8. Problem Statement Template format:
“We need to [WHAT – action/solution] for [WHO – stakeholders] in [WHERE – context] because [WHY – reason/impact].”
9. Vague problem made specific:
Vague: “Improve student learning.” Specific: “Help Class 8 students practice mathematical word problems through instant feedback on their solutions because many struggle with problem-solving and can’t get immediate teacher help.”
10. Why is the “Why” question important?
The “Why” question justifies resource investment, motivates stakeholders, helps prioritize efforts, and defines success metrics. Without a strong why, projects may lack support or direction.
F. Long Answer Questions – Answers
1. Explain the 4Ws Canvas with an example:
The 4Ws Canvas defines problems through four questions. For an AI lost-and-found system at school: WHO includes students who lose items, staff who manage lost items, and parents. WHAT is items getting lost and not being reunited with owners. WHERE is school premises including classrooms, corridors, and playground. WHY is students lose valuable items causing stress and parents spend money replacing them. This creates clarity for building an effective solution.
2. Why is Problem Scoping most important:
Problem Scoping is foundational because all subsequent stages depend on it. Poor scoping leads to collecting wrong data, building irrelevant models, and creating solutions nobody needs. It’s like building a house without architectural plans — you might end up with something unstable. Companies waste millions on AI projects that solve the wrong problems. Good scoping ensures everyone agrees on goals, success metrics exist, and resources are focused correctly.
3. 4Ws Canvas for study time management:
WHO: Students (especially those struggling with time), parents (concerned about study habits), teachers (assigning work). WHAT: Students don’t manage study time effectively, leading to last-minute cramming, incomplete homework, and stress. WHERE: Home study environment, across multiple subjects and assignments. WHY: Poor time management leads to lower grades, stress, and unhealthy study habits. Problem Statement: “We need an AI system that creates personalized study schedules for students at home because poor time management causes academic underperformance and stress.”
4. Theme, topic, and problem in education:
Theme: Education (very broad, includes everything about learning). Topic: Student assessment (more focused area within education). Problem: “Students don’t get immediate feedback on practice questions, so they repeat mistakes until the next test” (specific, actionable). The progression moves from general categories to specific issues that can be solved. An AI solution can address the specific problem but not the entire theme.
5. CottonAce case study:
The CottonAce app demonstrates successful problem scoping. WHO: Cotton farmers losing crops to pests. WHAT: Pink bollworm pest damages cotton, but farmers can’t identify it early. WHERE: Cotton fields across India during growing season. WHY: Farmers lose 20-30% of crops, spray pesticides blindly wasting money and harming environment. The clear problem led to an image-recognition AI solution. Result: 26.5% profit increase, 38% pesticide cost reduction. Good scoping ensured the solution addressed real farmer needs.
6. Common mistakes and how to avoid them:
Common mistakes include: Being too vague (fix: keep asking “what specifically?”), being too ambitious (fix: scope to achievable goals), solution-first thinking (fix: start with problems, not technology), missing stakeholders (fix: map all affected groups), and unclear success criteria (fix: define measurable outcomes). Prevent mistakes by always completing the 4Ws Canvas, validating problem understanding with stakeholders, and checking if the problem can realistically be solved with available resources.
7. 4Ws Canvas for canteen food wastage:
WHO: Students (waste food), canteen staff (throw away food), school management (lose money), environment (food in landfills). WHAT: 20-30 kg of food is thrown away daily because students take more than they eat. WHERE: School canteen during lunch hours, especially with popular but large portions. WHY: Food wastage costs money, contributes to environmental problems, and sets poor examples for students about sustainability. Problem Statement: “We need an AI system that predicts meal demand and recommends appropriate portion sizes for students in the school canteen because daily food wastage of 20-30 kg costs money and harms the environment.”
Activity Answer (Food Wastage Scenario)
Suggested 4Ws Canvas:
| W | Answer |
|---|---|
| WHO | Students (waste food), canteen staff (dispose waste), school management (financial loss), parents (pay for food), environment |
| WHAT | Students take more food than they can eat, resulting in 20-30 kg of daily food wastage |
| WHERE | School canteen during lunch breaks, particularly with buffet-style or fixed-portion meals |
| WHY | Wastes money (approximately ₹500-1000 daily), sends food to landfills, teaches poor sustainability habits, despite hunger existing elsewhere |
Problem Statement: “We need an AI system that monitors food consumption patterns and predicts appropriate portions for students in the school canteen because 20-30 kg of daily food waste costs money, harms the environment, and contradicts sustainability values.”
Previous Lesson: The AI Project Cycle Explained: 6 Steps to Build Your First AI Project
Next Lesson: Data Acquisition in AI: How to Collect, Source and Gather Data for Machine Learning Projects
