What the AI examiner does
Reads your thesis
Defensia indexes your uploaded thesis. The examiner draws on the full text to generate questions about your specific work — not a generic PhD question set.
Asks thesis-specific questions
Questions reference your actual methodology, your findings, your literature choices, and your arguments. They feel specific because they are.
Listens to your spoken answers
Your answer is transcribed in real time. The examiner uses what you said to decide what to ask next.
Follows up on your answers
Follow-up questions are generated from your responses. If your answer raises a further question, the examiner will ask it.
Gives structured feedback
After the session, you receive scores and notes across five categories: contribution, methodology, literature, critical thinking, and presentation.
Adapts to the style you choose
You choose the examiner style — supportive, rigorous, methodological, sceptical. The AI adapts its questioning to match.
Thesis-grounded questioning
The most important thing about Defensia is not that it uses AI — it's that the AI has read your thesis. Generic AI chatbots can ask viva-style questions. What makes Defensia different is that those questions are grounded in what you actually wrote.
When the examiner asks about your sampling approach, it's because your methodology chapter describes a specific sampling approach. When it asks about a limitation, it's one you acknowledged in your thesis. When it pushes back on a claim, it's a claim your thesis makes.
That specificity is what makes the practice realistic. And it's why the platform is most useful with a full thesis upload, not just an abstract.
A real-time spoken session
Question, answer, follow-up, feedback — in a live AI viva conversation.
Methodologist
Examiner speaking
EXAMINER
"Your methodology chapter describes a thematic analysis approach. Can you explain why you chose inductive coding rather than applying an existing theoretical framework to your data?"
YOUR ANSWER
"I chose inductive coding because my research question was exploratory — I was examining patterns in participant experience that hadn't been categorised before, so applying a predetermined framework would have constrained the analysis…"
FOLLOW-UP
"How did you establish rigour in that process? Did you use member checking or inter-rater reliability measures?"
SESSION FEEDBACK
What the AI can and can't do
Defensia is a preparation tool. It gives you a private space to practise spoken articulation, explore likely question areas, and rehearse follow-up conversations. It does not guarantee how your actual viva will go.
The AI examiner is not a human expert. It does not have the same intuition a specialist examiner brings, and it may occasionally ask questions that are off-target or not ideally phrased. That is worth knowing before you use it.
What it does reliably: ask specific questions from your thesis, listen to your answers, follow up on what you said, and give you structured practice of the spoken viva format. Most candidates find that sufficient to make preparation meaningfully better.
Defensia is a preparation tool. It is not a replacement for supervisors, formal mock vivas, or institutional guidance.