How Tutorwren works.
Every problem set is a diagnostic. We watch, detect, generate — and skip everything you already know.
How Tutorwren works — detailed steps
The problem set as a map
Every time a student attempts a question, Tutorwren records more than just right or wrong. It tracks time-on-task, error type, the sequence of attempts, and whether the student skipped and came back.
Each problem set becomes a live map of competency. Some skill nodes show mastery. Others show hesitation. A few show genuine breakdown. The AI builds this model in real time — no special setup, no teacher configuration required.
Stall detection
A stall isn't just a wrong answer. It's a pattern — spending 4× longer on one question type, making the same conceptual error twice in a row, skipping a problem and never returning, or checking the answer immediately instead of working through it.
Tutorwren analyzes these patterns to identify the exact skill node where understanding is fragile. Not "chapter 4 in general" — a specific gap like "equations where the variable appears on both sides."
Edge question generation
Once the stall node is identified, Tutorwren generates exactly three practice questions. Why three? Research on formative assessment suggests a short targeted practice set — not a full drill — is enough to consolidate understanding at a specific skill node. Each question is calibrated just past the point where mastery was last demonstrated.
These aren't pulled from a pre-set question bank. They're generated at the specific skill intersection — harder than what you showed you know, easier than what blocked you. The first question confirms the gap exists. The second applies it slightly differently. The third scaffolds toward the next concept. That sequence is intentional.
No wasted review
Traditional study: start from the beginning, re-highlight what you already know, slowly arrive at the hard part. By that point, attention is gone and time is spent. Tutorwren skips all of that.
When you've already mastered equations with one variable, you never see those again. When you've closed the gap on two-variable equations, the system moves to the next real edge. You spend your study time entirely on learning, not on re-learning.
Tutorwren vs. traditional study
| Dimension | Tutorwren | Traditional study |
|---|---|---|
| Time spent | 15–20 focused minutes on real gaps | 45–90 min, most on mastered material |
| What you practice | Exactly the skill node you're missing | Full chapter from the beginning |
| How much review | Zero repeat of mastered concepts | Majority of session is re-reading known content |
| Progress signal | Session data shows which skill nodes closed — not just "finished" | Completion feeling — doesn't distinguish mastery from exposure |