A right answer with no reasoning teaches nothing. Here's why every response from StudyHubs' tutor walks through the why, not just the what.
Most study tools optimize for the moment a student gets unstuck. Ours optimizes for the moment right after — when they hit a similar problem on their own and have to remember why the first one worked.
The answer isn't the point
It's tempting to measure a tutor by how fast it gets to a correct answer. But speed to answer is the wrong metric for learning — it's the right metric for cheating. We built the AI tutor around a slower, more deliberate loop: state the reasoning, show the step, then let the student try the next one themselves.
A student who memorizes an answer has solved one problem. A student who understands the reasoning has solved a category of problems.
In practice, that means every explanation is broken into small, checkable steps, and the tutor pauses to ask a student to predict the next one before revealing it. It's slightly slower than a flashcard app. It's also the reason our early classrooms saw retention numbers worth writing about.
What this looks like day to day
A tenth grader working through quadratic equations doesn't get 'x = 4' — they get the factoring step, a nudge toward why that factoring choice made sense, and a follow-up question that's structurally identical but numerically different. It costs us more compute per session. We think it's worth it.