Multiple choice is easy. Open-ended is the whole game — and the part parents dread. Here's how the app handles the long answers, and how the chat tutor digs deeper.
Feature B · 7 min read
Most "AI quiz" apps stop at multiple choice. That's not where the learning is. In Singapore schools, the marks that decide grades live in the open-ended sections — the "explain why", the "describe what happens", the "suggest a reason". Those are also the questions parents dread most, because reading twenty of them out loud, working out what the child meant, and writing helpful feedback takes an evening you don't have.
So this is the part of the app I cared about most.
When a student submits an open-ended answer, the AI doesn't just check if it matches a model answer. It walks through it the way a real teacher would:
After marking, the student gets back a small handful of things — not a wall of text:
The format is deliberate. A child who gets a one-line "wrong" learns nothing. A child who gets a specific, kind explanation of what they missed learns the concept.
Sometimes the feedback isn't enough. The student reads it and still doesn't get it. That's fine — and it's why the chat tutor exists.
After a marked answer, the student can hit the chat button and say whatever's on their mind:
The chat tutor has the question and the student's answer in front of it. It's not a generic chatbot — it knows what was just asked, what the student wrote, and where the gap is. It explains in plain language, gives a worked example if needed, and offers another practice problem if the student wants to try again.
The thing I cared about most: the chat never makes the student feel stupid for asking. There's no "as I already explained" energy. If they ask the same thing twice, it answers twice. That's the kind of patience a tired parent at 9pm doesn't always have, and the kind an expensive tutor usually does.
In my house, the open-ended questions were the ones that got skipped. Not because my kids didn't want to do them — because I didn't have time to mark twenty of them properly. So they'd do the multiple choice, leave the open-ended blank, and we'd both pretend that was fine.
That's the loop Professor Turtle breaks. The marking happens in seconds, with specific feedback. The chat tutor catches the "but I still don't get it" follow-up. So the hard questions actually get practised, and the child actually improves on the things that decide their grade.
It's not infallible. AI marking is a useful practice tool, not a definitive scorer. If a real exam were tomorrow and you wanted a final check on a borderline answer, ask a teacher, not an app. I've built this to help with the everyday grind of practice, not to replace professional judgement at the moment it matters most.