B10 · Beginner Capstone: AI for Social Good

Identifying a Real Problem

Identifying a real problem means finding a specific, evidenced need that affects real people, understanding its root cause, and describing it clearly in a problem statement, before thinking about technology. For the capstone, the problem should also be one where data and prediction could plausibly help, and one you can investigate yourself.

10 min Beginner Jump to the activity

What you need: Nothing. This page starts from the beginning, with no programming and no formulas.

Definition

Team Swachh's first idea was "an AI robot that cleans the whole town". Exciting, but is it the real problem? And could three students ever test it?

Identifying a real problem means finding a specific, evidenced need that affects real people, understanding its root cause, and describing it clearly in a problem statement, before thinking about technology. For the capstone, the problem should also be one where data and prediction could plausibly help, and one you can investigate yourself.

In simple words

Before a doctor treats you, they find out exactly what is wrong. Before you design an AI, you find out exactly what the problem is, who has it and why, by looking, asking and counting.

Why it matters

Most failed projects solve the wrong problem. Spending time here makes every later step easier and keeps your project useful to real people.

Key terms

Observation
Watching and recording what actually happens, with dates and details.
Interview
A planned conversation with someone affected, to learn about their experience.
Evidence
Facts, counts or quotes showing that a problem is real and how big it is.
Root cause
The underlying reason a problem happens.
Scope
The boundaries of a project: what it will and will not try to do.
Problem statement
One sentence naming who has the problem, what it is, where it happens and what an ideal solution would do.

Explained in detail

Gather evidence first

Team Swachh spent two weeks gathering evidence. Each morning on the way to school they recorded the state of six bins near the market: empty, half full, full or overflowing. They interviewed four shopkeepers, two residents and, with their teacher's help, a municipal sanitation supervisor. Their count showed that two of the six bins were overflowing on nine of the fourteen days, while the others were rarely more than half full. The supervisor explained that pickup trucks followed a fixed route on fixed days, with no way of knowing which bins were full.

Find the root cause

Asking "why" several times led the team from the symptom to the root cause. Why is there garbage on the street? Bins overflow. Why? Some bins fill much faster than others. Why are they not emptied? Trucks follow a fixed schedule. Why? The municipal team does not know which bins are full. That last answer is something information could fix, and it is far more specific than "the town is dirty". It also changed their idea: instead of a cleaning robot, they needed a way to tell the sanitation team which bins were full.

Team Swachh's 4Ws canvas

WTheir answer
WhoResidents and shopkeepers near the market; the municipal sanitation team
WhatTwo of six bins overflowed on 9 of 14 days while others were rarely half full
WhereSix public bins around the main market, especially after weekly market days
WhyLess street litter, fewer animals eating plastic, less wasted truck time

Their problem statement: "Residents and shopkeepers near our market have a problem that some public bins overflow for days while others are half empty, because the sanitation team cannot see which bins are full. An ideal solution would show the team, each morning, which bins need emptying."

Check the scope

  • Is it specific enough to test in a few weeks?
  • Do you have evidence, not just a feeling?
  • Could data and prediction genuinely help, or would a simpler fix work better?
  • Have you spoken to the people who have the problem?

Finally, check that you are allowed to investigate: ask a teacher before interviewing officials, and always explain to people why you are asking.

How it works

  1. Observe the problem and record evidence for at least a week.
  2. Interview at least three people affected.
  3. Ask "why" several times to find the root cause.
  4. Fill in the 4Ws canvas.
  5. Write your problem statement and check the scope.

Try it yourself

From robot to real problem

Team Swachh's first idea is a town-cleaning robot.

Evidence or assumption?

Sort each statement from a project journal.

  • The supervisor told us trucks follow a fixed weekly route

  • An app would definitely fix the problem

  • Bin 3 was overflowing on 9 of 14 mornings we checked

  • Everyone in town hates the bins

  • Four of five shopkeepers said smells were worst after market day

Common mistakes

  • Starting from a cool technology instead of a real need.
  • Relying on feelings instead of counted evidence.
  • Choosing a problem too big to investigate.

Check yourself

Question 1

1. What did Team Swachh's evidence reveal?

Question 2

2. Why did the team drop the robot idea?

Question 3

3. Which is evidence?

Summary

  • Gather evidence through observation and interviews before designing anything.
  • Ask "why" to reach the root cause, then fill in the 4Ws.
  • Write a specific, people-centred problem statement with a testable scope.

Finished reading?

Sources: CBSE Artificial Intelligence curriculum resources · Swachh Bharat Mission (Urban)