Analytics That Inform People — Not Scores That Label Students
Most "learning analytics" pitches lead with prediction: risk scores, dropout models, at-risk rankings. We build the other kind — descriptive dashboards and AI advisory flags on recorded data, surfaced to the staff whose judgment actually helps a student. What happened, clearly. What it means, decided by people.
What the analytics layer actually does
Descriptive dashboards on the single data model
Because attendance, enrollment, assessment records, and practice activity live on one platform, dashboards show coherent facts: attendance trends by cohort, submission patterns by course, practice-engagement aggregates by module. What happened, clearly — the foundation every intervention conversation starts from.
Advisory flags for staff review
AI agents watch the recorded data and flag patterns worth human attention: a student’s attendance shifting notably, a class’s practice results clustering on one concept, an anomalous assessment record. A flag is a prompt for a person to look — never a score, a ranking, or an automated action.
Aggregate practice insights for instructors
The practice quiz engine’s aggregate patterns show instructors which concepts a class finds hard while the course is still running. Cohort-level insight for teaching adjustments — individual practice results inform the student, not a profile.
Audit-logged, like all platform AI
Every advisory flag and AI interaction lands in the exportable audit log, and analytics access follows role-based permissions. Who saw what, and what the AI suggested, is always reviewable.
Why we deliberately don’t ship student risk prediction
- —Risk scores encode historical bias — students resembling past leavers get scored, not seen.
- —Predictions are unexplainable in the meeting that matters: with the student, the parent, or the board.
- —Regulation is moving against it: automated consequential profiling of students sits in the EU AI Act’s high-risk zone and draws growing scrutiny elsewhere.
- —The defensible pattern — a human reviewing a factual flag — catches the same situations without convicting anyone by statistics.
Related Resources
Frequently Asked Questions
Common questions from advisors, registrars, and institutional research teams.
See the dashboards and the flags — and what they don’t do
In a demo we will show the analytics on real workflows, including the audit trail — and you can hold the "no risk scores" claim against everything on screen.
Facts surfaced. Judgment human. Always.