Data Sources
Four inputs. One longitudinal picture per student.
The system ingests every meaningful signal the classroom generates. Each source adds a different dimension to the student’s profile. Together, they make it possible to say something specific and grounded about every individual.
Source 1
Discussion Transcripts
Anonymized, speaker-attributed records of every class session. The AI reads who spoke, how much, and about what — then flags engagement patterns a teacher can’t track while teaching.
Participation Rate
Concept Engagement
Peer Interaction
Depth of Response
Source 2
Tests & Quizzes
Structured assessments that map directly to course objectives. AI scores objective items, reads written responses, and connects results back to the student’s participation profile.
Mastery by Concept
Score Trends
Written Response Quality
Source 3
Worksheets & Written Work
In-class and take-home work products. Scanned and AI-read to extract evidence of reasoning, not just correctness — flagging where a student understands the answer but not yet the why.
Reasoning Quality
Effort Signals
Concept Application
Source 4
Project & Portfolio Work
Extended work products: research, presentations, prototypes, and creative submissions. Evaluated against rubrics and connected to discussion behavior — does what they build reflect what they say in class?
Rubric Alignment
Growth Over Time
Talk-to-Build Coherence
Intelligence Out
Not dashboards. Advice.
The system doesn’t present data for teachers to interpret. It interprets the data and delivers specific, actionable guidance — personalized to each student, grounded in evidence from that student’s own work.
For the Teacher
Per-Student Coaching Prompts
Before each class, the teacher receives a brief on every student: what they showed this week, what’s missing, and a suggested approach. Not a summary — a next step.
Example prompt
“Student 4 can restate the Five Pillars correctly but hasn’t yet applied them to an original scenario. The quiz confirms it. Push for application today.”
For the Parent
Specific, Grounded Updates
Parents receive a teacher-approved summary rooted in actual classroom behavior — not a grade and a generic comment. Every note traces back to something the student said, wrote, or built.
Example note
“Your child asked the most probing question of Session 11 — about whether intent changes ethical responsibility. That’s a college-level distinction. They’re thinking at that level.”
For the School
Aggregate Pattern Detection
Administrators see anonymized trends across the whole class: which concepts are landing, where engagement falls, and how the course is developing over time. No individual student is identifiable.
Example signal
“Engagement dropped 18% in Sessions 8–10. The transcript shows the discussion format changed. Consider returning to structured Socratic dialogue.”
Intervention Alerts
Catch the Slide Before It’s a Problem
The system monitors longitudinal patterns, not just individual sessions. When a student’s participation drops, their work quality slips, or a disconnect emerges between what they say and what they submit, the teacher receives a flag — not after the unit ends, but during it.
Pre-Class Briefs
Walk In Already Prepared
Every session begins with a one-page AI brief: what carried over from last time, who may need a nudge, and suggested anchors for discussion. The teacher reads it, adjusts if needed, and teaches. The brief writes itself from the data that already exists.
The Data Flow
Captured once. Useful everywhere.
Every data input feeds a shared student profile. That profile is the source of truth for every stakeholder view — teacher, student, parent, admin — so the same underlying evidence reaches every person in the form they need.
1
Data enters from any source
Transcript, assessment, worksheet, or project — each is tagged to the student code and ingested.
2
AI reads and scores it
Engagement, mastery, reasoning quality, and behavioral signals are extracted and written to the student’s longitudinal profile.
3
Teacher reviews and approves
Nothing reaches a student or parent without passing through a teacher gate. AI generates. The teacher decides.
4
Each stakeholder gets their view
The same underlying data surfaces as coaching prompts for the teacher, summaries for parents, a profile for the student, and aggregate trends for admin.
Student codes, not names
Every student is a randomly generated code in the system. Real identity is stored separately, encrypted, and never joins with behavioral or academic data during normal operation.
Role-bounded by design
Each stakeholder view is enforced at the database level, not just the interface. A parent cannot access another student’s data even with a direct API call. The boundary is structural.
Every access is logged
All data queries are recorded: who, what, when, and from which role. Administrators can pull a full access log at any time. Nothing happens without a record.
Stakeholders
The right information for every person in the room.
Six roles. Each gets a view built for their relationship to the student — not a filtered version of someone else’s dashboard.
Teacher
Coaching Prompts & Pre-Class Briefs
Per-student coaching cards, intervention alerts, lesson planning workspace, and a nightly brief before each session. Full access to real names and raw data. Controls what every other role sees.
Student
Their Own Profile
Sees their engagement history, project status, and assessment results. When the teacher enables it, an AI tutor answers questions scoped to course content. Never sees other students.
Parent
Specific, Approved Updates
Receives a teacher-approved weekly digest: what their child demonstrated, where they showed growth, what to ask about. Grounded in actual classroom evidence, not generic progress notes.
Administrator
Anonymized Aggregate Trends
Sees class-wide engagement patterns, concept mastery distribution, and session quality over time. No individual student is ever identifiable in the admin view.
Dept. Lead
Cross-Class Intelligence
Anonymized analytics across multiple teachers’ sections — identifying curriculum gaps and engagement patterns at the department level without accessing any individual classroom data.
IT / System Admin
Infrastructure & Compliance
Manages authentication, security rules, and cost monitoring. High setup effort; low steady-state overhead. Full audit log available for compliance reporting at any time.