Use Case: ENEM Prep School with Studeia

The ENEM Prep School Problem

ENEM prep schools (cursinhos — Brazil's university-entrance prep academies) face 4 structural challenges:

1. High dropout — 30–50% of students quit before the ENEM exam (motivation, cost, difficulty)
2. Manual mock exams drain time — preparing a 90-question mock exam plus grading plus analytics takes weeks
3. In-person tutoring is expensive — answering individual questions around the clock would require a large teaching staff
4. Parents want visibility — especially for minors, parents demand progress updates but schools don't have the bandwidth

Studeia addresses all 4.

How Studeia Solves It

1. Automated mock exams with psychometric analytics

Configure a mock exam once, reuse it for multiple class groups:

{
  "configuration": {
    "maxAttempts": 1,
    "timeLimitMinutes": 270,
    "shuffleQuestions": true,
    "shuffleOptions": true,
    "scoringPolicy": "highest",
    "reviewMode": "after_due_date",
    "passingScore": 50
  },
  "questions": [...]  // or import from question_bank
}

After the mock exam, analytics are available at /institution/courses/[id]/lessons/[lid]/quiz-analytics:

• Item difficulty — which questions 80% of students get right (too easy) or fewer than 30% get right (too hard)
• Point-biserial discrimination — flags items that do NOT differentiate strong from weak students (candidates for removal)
• Distractor analysis — which wrong answer is the most popular (signals a specific misconception)
• Cronbach's alpha — internal consistency of the exam (should be > 0.7 for reliability)
• Score distribution — score histogram (detects ceiling/floor effects)

2. Importable question bank

Prep schools already have question banks in PDF, Word, or spreadsheet form. Studeia accepts:

• GIFT format (standard Moodle/Canvas format, fully compatible)
• CSV (one question per row, answer choices in columns)
• API JSON (POST /api/institution/question-banks/[id]/items)
• Manual via UI (visual editor)

Each question has concepts[] (mapped to a student's ConceptMastery). When the AI tutor detects a misconception linked to a concept, it suggests re-attempting related questions.

3. Personalized AI tutor for each student

School materials are ingested into the RAG (POST /api/institution/courses/[id]/rag-ingest):

• PDF course booklets
• Lecture videos with auto-transcription (Whisper)
• Slides
• Worked solutions

When a student asks "How do I solve question 47 from the Math mock exam?":

1. RetrievalAgent fetches the worked solution from the materials
2. PedagogicalAgent selects a strategy based on the student's ConceptMastery in Math
3. The tutor responds citing the specific booklet or lecture
4. EvaluationAgent detects misconceptions (e.g., "student confuses domain vs. range")
5. Proactive inline quiz: "Let me ask you 2 similar questions to confirm you've got it"

4. Gamification to fight dropout

XP per activity plus streak bonuses:

Day 1:  50 XP (lesson) + 30 XP (quiz)
Day 2:  50 XP + 30 XP + 2-day streak bonus
Day 7:  50 XP + 30 XP + 7-day streak bonus (+100 XP)
Day 30: 50 XP + 30 XP + 30-day streak bonus (+500 XP)

Badges with 11 automatic criteria:

• "Marathoner" — 30-day streak
• "Math Machine" — 100 correct Math answers
• "Top 10%" — in the top 10% on a mock exam
• Custom badges created by admins via the AI gamification agent

Class-scoped leaderboards with weekly windows sustain healthy competition without a global ranking (which demotivates lower-performing students).

Timed contests (e.g., "Weekly Championship — who completes the most quizzes"):

• Prize: special badge + 1,000 coins
• Students redeem coins for custom avatars, certificate templates, and more

5. AI-powered parent portal

GuardianLink verified by SMS (LGPD + ECA compliant for minors):

• Link 1 parent → multiple children
• Parent chooses access level per child (Progress, Chat history, Time restrictions)
• Automatic alerts via push/email/in-app:
  • Inactivity (>5 days without login)
  • Performance drop (>20% vs. historical average in a given week)
  • Milestone reached (module completed, special badge earned)
  • Difficulty with a topic (mastery < 0.4 on a key concept)

Periodic reports generated by LLM:

• Daily (Haiku): "João studied for 45 min today and completed 3 Chemistry lessons."
• Weekly (Sonnet): "This week João made progress in Chemistry (mastery +0.15) but regressed in Math..."
• Monthly (Sonnet with charts): full pedagogical narrative + recommendations

Typical Setup (1 Week)

TCO — 200 Students / 3 Years

Studeia is ~24% cheaper plus 1-week setup vs. 60–90 days for Moodle.

From Plan to Practice

The 1-week setup above requires no contract: start on the free Demo plan, import a real question bank, and run a pilot mock exam with one class group. If the ENEM calendar is already running, every week you delay is one fewer week of mock-exam data to calibrate your students' preparation. See pricing and details on the prep school page.

Transparency note: the scenarios and TCO figures on this page are illustrative estimates, not real customer cases — run your own numbers using the public pricing.

See Also

• Quiz Engine
• RAG Ingestion
• Gamification
• Studeia vs. Moodle Comparison