How My Son Got His First 90s — And What Actually Made the Difference
I'm not an edtech person. I'm a dad who built a scrappy AI tutor for his kid. Here's exactly what we did, what didn't work, and why two exams later he came home with 90s.
I want to be upfront about something: this isn’t a product pitch disguised as a blog post.
This is just the story of what happened when I tried to help my 10-year-old son prep for two exams — and ended up accidentally building GradeQuests in the process.
The problem with worksheets
My son shuts down the second a worksheet appears. It doesn’t matter if it’s math or language arts. The moment he sees a sheet of printed questions, something in him switches off. Eyes glaze. Pencil disappears. An inexplicable need to reorganize his Lego collection emerges.
I’ve tried sitting next to him. I’ve tried timers. I’ve tried bribery (it works for about four minutes). The issue isn’t that he’s not smart — he is. The issue is that worksheets feel like a punishment.
What I built instead
I’m a developer. So I did what developers do: I over-engineered a solution to a parenting problem.
The first version was embarrassingly simple. I’d take a photo of whatever he was working on, feed it to a language model, and have it generate 5 questions based on the content. Then I’d paste those questions into a basic web page with multiple choice answers.
That’s it. That was the whole thing.
The first session
He sat down expecting worksheets. Instead I handed him the iPad with a simple question and four answer choices.
“What is this?”
“Just answer the question.”
He got it wrong. The app explained why — short, simple, no lectures. Then the next question appeared.
He played for 22 minutes without me having to say a word.
The exams
We did two rounds of sessions before each exam. Nothing fancy — 15 to 20 minutes each, usually in the evening. I’d snap a photo of the unit review sheet, generate the quiz, and leave him to it.
First exam: Math. He got a 90.
I honestly didn’t believe it at first. His previous math mark was 68.
Second exam, a few weeks later: Language Arts. 91.
He came home and asked if we could do another quest that evening — not because I asked him to.
What I think actually made the difference
I’ve thought about this a lot. A few things stand out:
Short sessions beat long ones. 15-20 minutes of focused practice is more valuable than an hour of half-attention. The format forced shorter sessions.
Immediate feedback matters. He found out instantly whether he was right or wrong — and why. No waiting for a corrected worksheet two days later.
The game framing removed the stakes. Calling it a “quest” and giving him XP made it feel less like a test and more like practice. The pressure was off.
I wasn’t in the room. This surprised me most. When I sat beside him, he looked to me for validation on every question. On his own, he actually had to think.
What didn’t work
I tried making the sessions longer. He lost focus after about 20 minutes consistently.
I tried covering material he hadn’t learned yet. That backfired — he got frustrated and disengaged. The app works best as review, not introduction.
I also tried using it every single day for a week straight. By day four, the novelty had worn off. Three or four times a week felt better.
Why I’m building this into a product
After the second 90, a few parents at school asked what we were doing differently. I tried to explain. It’s clunky to explain.
So I’m cleaning it up and making it available to other parents. Not because I think it’s magic — it isn’t. But because the core loop (photo → short quiz → immediate feedback → XP) actually works, and I want other kids to have access to it.
If your kid gives it a fair shot on one worksheet and doesn’t engage, don’t bother. But in my experience — and I only have one data point, my son — the format is genuinely different enough that kids who’ve checked out of traditional studying will check back in.
GradeQuests is available to try free. No card required — takes about 2 minutes to get started.