Obsidian Intelligence integrates observable performance data with neuroscience-informed principles of attention, memory, cognitive load, and retrieval. It is not an AI tutor. It is a score-conversion system.
Obsidian maps the student’s current score profile beyond broad sections. It looks at question type, difficulty tolerance, pace, answer changes, fatigue patterns, and the distinction between knowledge gaps and execution failures.
Patterns by domain, sub-skill, difficulty band, and point in the module.
Where the student rushes, hesitates, over-invests, or loses stability under time pressure.
A missed question is not a diagnosis. Obsidian classifies whether the cause was content, translation, retrieval, attention, pacing, strategy, or a pressure-based decision error. The purpose is to stop wasting preparation time on the wrong intervention.
Performance in the first half of each SAT section is treated as a strategic bottleneck, not a generic warm-up.
Students receive concrete rules for skipping, returning, checking, and changing answers.
Practice is continuously redesigned around the student’s live profile. The system decides what to drill, what to stop drilling, what to revisit later, and when the student needs simulation rather than more content work.
The output is not a generic calendar. It is a private weekly protocol intended to recover the highest-leverage points first.