Student Productivity
How to Defend a Project You Built With AI in Your Viva
Using AI to build your project is not the problem — not understanding it is. A preparation plan for defending AI-assisted work honestly in your viva.
If you built your final-year project or semester project with AI help, the viva can feel like a trap. You are proud of what you built, but somewhere in the back of your mind sits a question: what if the examiner asks about a part I do not fully understand? That fear is reasonable, and it is also fixable. A viva does not test whether you typed every character yourself. It tests whether you understand what you built, why you built it that way, and what you would do differently next time. Those are things you can prepare for, even if AI helped you write the code.
The first thing to accept is that using AI is not, by itself, a problem. Examiners know these tools exist, and many of them use similar tools in their own work. What they will not accept is a student who cannot explain their own project. The difference between a strong viva and a painful one is rarely the tool you used. It is the depth of your understanding. This post is a preparation plan for exactly that.
Start with honesty about what AI did
If your examiner asks whether you used AI, say yes and be specific. Vague answers sound like hiding. A good answer sounds like this: I used AI to scaffold the project structure and to draft the database queries, but I designed the features myself, I debugged the authentication flow myself, and I rewrote the parts that did not fit. That answer does three things at once: it is honest, it shows you know which parts are yours, and it proves you stayed in charge of the decisions.
Hiding AI use is the one move that reliably backfires. If an examiner finds AI-generated patterns in your code and you have denied using AI, the conversation stops being about your project and starts being about your honesty. If you are open about it from the start, the conversation stays where you want it: on your understanding.
Know every part of the project, not just the demo path
Most students prepare the happy path: the features that work when everything goes right. Examiners deliberately walk off that path. They click the button you never tested. They ask about the file you copied in and never read. So before your viva, walk through your own project like an examiner would. For every module, be able to answer four questions:
- What does this part do? Say it in one plain sentence, without jargon.
- Why did you build it this way? Name the alternative you rejected and why.
- What breaks it? Every feature has an edge case. Know yours.
- What would you change? A thoughtful improvement beats a claim of perfection.
The fastest way to find your gaps is the explain-it-to-a-junior test. Pick a classmate who has not seen your project and explain each feature to them in plain words, without opening the code. Every time you get stuck, you have found a gap to close before the viva, not during it.
Build a decisions log before the viva
Examiners love decisions, because decisions are where thinking shows. Before your viva, write down eight to ten decisions you made during the project, each with a reason. They do not need to be dramatic. Choosing one database over another because your app works offline is a decision. Switching a library halfway because the first one made your code unreadable is a decision. Dropping a feature because the deadline forced a trade-off is a decision, and a mature one.
Keep this log short: one line for the choice, one line for the reason, one line for the alternative you considered. Read it the night before your viva. When the examiner asks why you did something, you will not be improvising. You will be recalling work you already did. If you want a structure for presenting those decisions, the case-study format in How to Turn GitHub Repos Into Portfolio Case Studies works just as well in a viva as it does on a portfolio page.
Prepare for the questions examiners actually ask
Viva questions are more predictable than students expect. Practice answering these out loud, not just in your head:
- Why did you choose this technology? Answer with a trade-off, not a slogan. Never say you chose something because it is the best. Say what it gave you and what it cost.
- What was the hardest part? Pick something real. Examiners can tell the difference between a rehearsed struggle and a genuine one.
- What happens when this fails? Errors, empty states, bad input, no internet. If you have never thought about failure, think about it now.
- Show me the part you understand least. This is a trap only if you bluff. Naming a weak spot and explaining how you would learn it is a strong answer.
- What would you add with one more month? This tests judgment. Pick the improvement with the highest value, and say why.
- Did you write this code yourself? Answer honestly, name what AI drafted and what you designed, debugged, or rewrote.
The demo is your best defense
A working demo changes the whole tone of a viva. It moves the conversation from abstract doubt to concrete evidence. Rehearse the exact path you will show: the feature that works best, in the order that tells the story of the project. Keep it short. Three minutes of a smooth demo beats fifteen minutes of clicking around hoping something loads.
And prepare for the demo to break, because demos love breaking at the worst moment. Have a fallback ready: screenshots of the key screens, or a short screen recording. If something fails live, stay calm, explain what should have happened, and show the fallback. Examiners remember composure under a broken demo more than they remember the break itself.
What to say when you genuinely do not know
You will be asked something you cannot answer. Every student is. The worst response is a confident guess dressed up as knowledge, because examiners usually know the answer and are testing whether you do. The best response is honest and forward-looking: I do not know that part well, but here is how I would find out, and here is what I do understand that is adjacent to it. That answer protects your credibility and often earns you a follow-up question you can answer.
A viva is not an interrogation about your tools. It is a conversation about your thinking — and thinking is the one part AI cannot do for you.
A one-week viva preparation plan
If your viva is a week away, here is a plan that covers everything above without panic:
- Day 1–2: Walk the whole codebase. Write your decisions log. List every feature and its edge cases.
- Day 3: Explain the project to a classmate without opening the code. Note every gap and close it.
- Day 4: Rehearse the demo twice, then deliberately break it once and practice your recovery.
- Day 5: Practice the six hard questions out loud. Record yourself if you can stand it — it helps.
- Day 6: Read your own report or README end to end. Examiners often ask about lines students forgot they wrote.
- Day 7: Rest. Skim the decisions log once. A rested brain defends better than a crammed one.
You built the project. The AI was a tool, not the author of your understanding. Walk in ready to show the thinking behind the code, and the viva becomes what it is supposed to be: a chance to prove you know your own work. For help turning that same project into proof after the viva, see how ZEVQYN works — the same chain from research to portfolio applies to viva preparation too.