Five prioritized improvements for AI tutoring prototypes: instant core-strategy layer, curriculum tags, dynamic geometry, step-specific buttons, automated wrong-book¶

Strategy · All strategies
Evidence · 4 claims (4 for) · 4 studies (4 design), q2 · 0 of 4 report an effect size · 1 claim rests on one study

Description¶

Based on field findings and participants' explicit requests, the authors list five improvements: a dual-mode structure where the "Answer + Core Idea" layer loads instantly; explicit junior-high syllabus tags; automatic highlighting of auxiliary lines synchronized with text; localized step-specific follow-up buttons such as "Explain this step"; and automated wrong-book generation organizing captured problems by knowledge point for spaced weekend practice.

Design Implications¶

Context¶

Requirements¶

• telemetry or observational evidence identifying the corresponding breakdown before prioritizing an improvement

Constraints¶

• derived from a 12-day, 12-participant deployment; the authors note it does not support population-level claims

Target Learners¶

• junior-high students under homework time pressure

Target Learning Goals¶

• reducing abandonment, sustaining curriculum-valid methods, enabling spaced retrieval

Claims¶

• In high-stakes homework settings, answer-first checking functions as a rational self-regulation strategy and diagnostic checkpoint rather than shortcutting [+M]
• For exam-oriented learners, trust in AI tutors hinges on curriculum fit: students abandon mathematically correct solutions that use out-of-syllabus methods [+M]
• Step-linked visual grounding externalizes spatial reasoning: synchronized diagrams that reveal auxiliary lines step by step reduce split-attention effort [+M]
• Students rarely open transfer-practice cards immediately after solving; they want such problems organized into delayed, spaced wrong-book review instead [+M]

Related Strategies¶

Examples¶

-

Key Sources¶

• Harry Feng, Yuan Tian, and Erica Zhao. 2026. From Answer Generators to Reasoning Facilitators: Designing AI Tutors for Mathematical Reasoning in High-Stakes Environments. https://arxiv.org/abs/2607.01692