Definition
You have learned what AI is, how it learns, how to use it safely and how a project moves from problem to deployment. Now it is your turn: choose a real problem around you and design an AI idea to help.
The capstone is a final project that brings together everything in the Beginner level. Working alone or in a small team, you will identify a real problem in your community, design an AI-based solution using the AI project cycle, build a simple prototype with a no-code tool, review it for fairness and safety, and present a proposal. You are not expected to build a finished product; you are expected to think it through carefully and honestly.
In simple words
It is like planning a science fair project, but about AI and a real need. You find something that bothers people, imagine how AI could help, make a small working model, check it is fair and safe, and explain your idea to others.
Why it matters
Applying ideas to a real problem is the best way to understand them deeply. The capstone also gives you something to show, a project that demonstrates how you think, which matters more than any certificate.
Key terms
- Capstone
- A final project that combines what has been learned in a course.
- Deliverable
- Something you must produce and hand in, such as a proposal or a presentation.
- Milestone
- A checkpoint in a project, such as finishing the problem statement.
- Prototype
- A simple early version that shows how an idea could work.
- Proposal
- A document that explains a problem, a solution and a plan, to convince others it is worth doing.
- Reflection
- Thinking back on what you did, what worked, what did not and what you learned.
Explained in detail
The eight steps
| Step | What you produce | Links to |
|---|---|---|
| 1. Project briefing | A team, a timeline and a project journal | This page |
| 2. Identify a real problem | A 4Ws canvas and problem statement | Problem scoping |
| 3. Users, benefits and risks | A stakeholder map and risk list | Design thinking, ethics |
| 4. Plan the data | A system map and data plan | Data acquisition and exploration |
| 5. Build a prototype | A working no-code model and test results | Modelling, evaluation |
| 6. Fairness and safety review | A checklist with findings and fixes | Responsible AI |
| 7. Create the proposal | A short written proposal | Data storytelling |
| 8. Present and reflect | A presentation and a reflection | Communication |
Our running example: Team Swachh
Throughout this module, you will follow Team Swachh, three students in a small town who noticed that public garbage bins near their market often overflow for days, attracting stray animals and spreading smells, while other bins are emptied when half full. Their idea: residents send a photo of a bin, an AI model classifies it as "overflowing" or "fine", and the municipal team uses a daily map of overflowing bins to plan pickups. Each page shows how the team handled that step, including their mistakes, and then asks you to do the same for your own project.
Choosing a project area
- Environment: waste, water, trees, air quality, animals.
- Health: reminders for medicines, healthy food in the canteen, first-aid information.
- Education: helping younger students read, finding lost library books, learning in local languages.
- Safety: road crossings, scam awareness for elders, reporting broken streetlights.
- Inclusion: tools for people with disabilities, help for people who cannot read.
Working well
Keep a project journal from day one: a running record, on paper or in a shared document, of decisions, evidence, problems and changes. Divide work fairly, meet regularly and agree how decisions are made. Set milestones for each step. Above all, remember that the goal is careful, honest thinking. A modest idea, well tested and honestly described, is worth more than an impressive idea with no evidence.
You will not need to write any programs for this capstone. Everything can be done with observation, interviews, paper, a spreadsheet and a free no-code tool in a web browser. What you do need is curiosity, patience and honesty.
How it works
- Form a team of two to four, or decide to work alone.
- Start a project journal.
- List three or four problem areas that interest you.
- Agree on a timeline with a milestone for each of the eight steps.
- Decide who will do what, and how you will make decisions.
Try it yourself
How Team Swachh got started
Three friends walk past the market every day on their way to school.
The capstone in order
Put the capstone steps in order.
- 1Identify a real problem
- 2Study users, benefits and risks
- 3Review fairness and safety
- 4Plan the data
- 5Write the proposal and present it
- 6Build a prototype
Common mistakes
- Choosing a problem you cannot observe or people you cannot reach.
- Skipping the project journal and losing track of decisions.
- Aiming for an impressive idea instead of a well-tested, honest one.
Check yourself
Summary
- The capstone applies everything from the Beginner level to a real problem.
- It has eight steps, following the AI project cycle with ethics built in.
- Keep a journal, plan milestones and value honest, well-tested thinking.
Finished reading?
Sources: United Nations Sustainable Development Goals · CBSE Artificial Intelligence curriculum resources