AI-Powered Insights Guide | Documentation

1. Overview

DojoMaster's AI-powered insights use advanced machine learning to analyze student data and provide actionable predictions. These insights help instructors identify at-risk students, determine grading readiness, and receive personalized recommendations for each student's development.

Key Features

Retention Risk Prediction: Identify students at risk of dropping out with 4 risk levels (LOW, MEDIUM, HIGH, CRITICAL)
Grading Readiness Assessment: Determine which students are ready for their next belt with 5 readiness levels
Comprehensive Student Insights: Receive holistic analysis including strengths, areas for improvement, and personalized recommendations
Progress Trajectory Forecasting: See projected student development over time
Club-Wide Analytics: View aggregated insights across all students to identify trends
Intelligent Caching: Insights are cached for 24 hours to optimize performance and reduce AI costs
Confidence Scoring: Each prediction includes a confidence score (0-100) indicating reliability

AI-Enabled Subscription Required

AI-powered insights require the AI-Enabled subscription tier (A$199/month). This tier includes unlimited students, all Standard features, plus AI-powered predictions and analytics. Upgrade in Dashboard → Billing.

2. Prerequisites

Before using AI-powered insights, ensure you have:

✓OWNER or SENSEI role (required to access AI insights)
✓AI-Enabled subscription tier (A$199/month with active subscription status)
✓Student data (students enrolled with attendance, belt ranks, and stripe progress)
✓Historical data (at least 2-3 months of attendance and progress data for accurate predictions)
✓AI features enabled (globally enabled in system configuration)

3. Understanding AI Insights

AI insights analyze multiple data points to provide comprehensive predictions about student performance and engagement.

How AI Insights Work

1.Data Aggregation: System collects student data including attendance, belt progress, stripe awards, grading history, and skill proficiency
2.AI Analysis: OpenAI's GPT-4 model analyzes the data using structured prompts designed for martial arts education
3.Prediction Generation: AI generates predictions for retention risk, grading readiness, and comprehensive insights
4.Caching: Insights are saved to the database and cached for 24 hours to optimize performance
5.Display: Insights appear in the AI Insights Dashboard and individual student profiles

Data Sources for AI Predictions

•Attendance Records: Attendance rate, recent attendance trend, days since last attendance, current streak
•Belt Progress: Current belt rank, months since last grading, belt progression history
•Stripe Awards: Stripes earned, unique stripe types earned, stripe completion percentage
•Grading History: Past grading results, scores, feedback, assessor notes
•Skill Proficiency: Required skills, mastered skills, average skill proficiency score
•Training Duration: Days trained, training days required for next belt, training consistency

4. Types of AI Predictions

DojoMaster provides three types of AI-powered predictions:

Retention Risk Prediction

Identifies students at risk of dropping out based on attendance patterns, progress stagnation, and engagement levels.

🟢 LOW Risk (0-25)

Student is highly engaged with consistent attendance, regular progress, and strong skill development. No intervention needed.

🟡 MEDIUM Risk (26-50)

Student shows some warning signs such as declining attendance or slower progress. Monitor closely and provide encouragement.

🟠 HIGH Risk (51-75)

Student has significant risk factors such as frequent absences or no recent progress. Proactive intervention recommended.

🔴 CRITICAL Risk (76-100)

Student is at very high risk of dropping out with multiple concerning factors. Immediate intervention required.

Grading Readiness Assessment

Evaluates whether students are ready for their next belt grading based on attendance, training days, stripe completion, and skill mastery.

❌ NOT_READY (0-20)

Student does not meet basic requirements. Significant training and progress needed before grading.

⚠️ NEEDS_IMPROVEMENT (21-50)

Student meets some requirements but has gaps in attendance, stripes, or skills. More preparation needed.

🟡 ALMOST_READY (51-75)

Student is close to ready with most requirements met. A few more weeks of training recommended.

✅ READY (76-90)

Student meets all requirements and is ready for grading. Can be enrolled in next grading event.

⭐ HIGHLY_READY (91-100)

Student exceeds all requirements with exceptional performance. Highly likely to pass grading.

Comprehensive Student Insights

Provides holistic analysis combining retention risk and grading readiness with personalized recommendations.

•Overall Performance Summary: High-level assessment of student's current status
•Key Strengths: Areas where the student excels (e.g., consistent attendance, skill mastery)
•Areas for Improvement: Specific aspects needing attention (e.g., attendance consistency, specific skills)
•Personalized Recommendations: Actionable suggestions for instructors and students
•Motivational Message: Encouraging message tailored to the student's progress
•Progress Trajectory: Current status, projected development, and estimated timeframe

5. Viewing AI Insights

AI insights are accessible from multiple locations in DojoMaster:

AI Insights Dashboard

1.Navigate to Dashboard → AI Insights
2.View club-wide analytics with five tabs:

📊 Overview Tab

Club-wide statistics including total students analyzed, average retention risk, students ready for grading, and quick actions.

⚠️ At-Risk Students Tab

List of students with MEDIUM, HIGH, or CRITICAL retention risk. Shows risk level, risk score, key factors, and recommendations.

✅ Ready for Grading Tab

Students with READY or HIGHLY_READY grading readiness. Shows readiness level, target belt, and estimated time to readiness.

⭐ Top Performers Tab

Students with exceptional performance (LOW retention risk + READY/HIGHLY_READY grading readiness). Celebrate their achievements!

📋 All Insights Tab

Complete list of all student insights with retention risk, grading readiness, and last generated timestamp.

Individual Student Profile

1.Navigate to Students → [Student Name]
2.Scroll to the "AI Insights" section
3.View comprehensive insights including retention risk, grading readiness, strengths, areas for improvement, and recommendations
4.Click "Analyze Student with AI" to generate fresh insights

6. Generating AI Insights

AI insights can be generated for individual students or all students at once.

Generate Insights for One Student

1.Open the student's profile page
2.Scroll to the AI Insights section
3.Click "Analyze Student with AI" button
4.Wait 5-10 seconds for AI analysis to complete
5.View the generated insights including retention risk, grading readiness, and recommendations

Generate Insights for All Students

1.Navigate to Dashboard → AI Insights
2.Click "Analyze All Students" button in the top-right corner
3.Wait for the analysis to complete (time varies based on number of students)
4.View club-wide analytics and individual student insights

Tip: Intelligent Caching

AI insights are cached for 24 hours. If you generate insights for a student and then view them again within 24 hours, the cached version is displayed instantly. After 24 hours, insights are marked as "stale" and you can regenerate them for updated predictions.

7. Interpreting Predictions

Understanding how to read and act on AI predictions is crucial for effective student management.

Understanding Confidence Scores

Each prediction includes a confidence score (0-100) indicating how reliable the AI considers its prediction:

•90-100: Very high confidence - prediction is highly reliable based on strong data patterns
•70-89: High confidence - prediction is reliable with good supporting data
•50-69: Moderate confidence - prediction is reasonable but consider other factors
•Below 50: Low confidence - limited data available, use prediction as one of many factors

Acting on Retention Risk Predictions

🟢 LOW Risk

Action: Continue current approach. Provide positive reinforcement and recognition for their commitment.

🟡 MEDIUM Risk

Action: Monitor closely. Have a casual conversation to check in on their experience. Provide extra encouragement and support.

🟠 HIGH Risk

Action: Proactive intervention needed. Schedule a one-on-one meeting to discuss challenges. Offer flexible scheduling or additional support.

🔴 CRITICAL Risk

Action: Immediate intervention required. Contact student/parent urgently. Identify barriers and create action plan to re-engage.

Acting on Grading Readiness Predictions

•HIGHLY_READY / READY: Enroll in next grading event. Provide final preparation and encouragement.
•ALMOST_READY: Review specific areas needing improvement. Set clear goals for next 2-4 weeks.
•NEEDS_IMPROVEMENT: Focus on identified gaps. Provide targeted training and feedback.
•NOT_READY: Continue regular training. Set realistic timeline for grading readiness (typically 3-6 months).

8. Privacy and Data Usage

DojoMaster takes student privacy seriously and follows strict data protection practices for AI features.

How Student Data is Used

•Data Aggregation: Only aggregated, anonymized data is sent to OpenAI for analysis (no personally identifiable information)
•Minimal Data Sharing: Only essential metrics are shared (attendance rates, belt progress, stripe counts) - not names, addresses, or contact information
•Secure Storage: All AI insights are stored securely in DojoMaster's database with encryption
•Access Control: Only OWNER and SENSEI roles can view AI insights
•No Training Data: OpenAI does not use DojoMaster data to train their models (per API agreement)

Student Access to AI Insights

Students can view their own AI insights through the Student Portal:

•Students see motivational messages and progress trajectory
•Students do NOT see retention risk scores or detailed instructor recommendations
•Insights are presented in an encouraging, positive manner
•Parents can view their child's insights through the Parent Portal

Compliance and Regulations

•GDPR Compliant: AI features comply with GDPR data protection requirements
•Australian Privacy Principles: Adheres to APPs for Australian clubs
•Data Retention: AI insights are retained for 12 months, then automatically deleted
•Right to Deletion: Students/parents can request deletion of AI insights at any time

Privacy Commitment

DojoMaster never sells student data to third parties. AI features are designed to enhance student experience while maintaining the highest privacy standards. For full details, see our Privacy Policy and Security Policy.

9. How AI Models Work

Understanding how AI generates predictions helps you interpret and trust the insights.

AI Technology Stack

•Model: OpenAI GPT-4 (latest version) with structured output
•Temperature: 0.3 for predictions (low randomness, high consistency), 0.5 for insights (slightly more creative)
•Max Tokens: 1000-1500 tokens per prediction (ensures detailed responses)
•System Prompt: Custom prompt designed for martial arts education context

Prediction Accuracy

AI prediction accuracy depends on several factors:

•Data Quality: More complete and accurate data leads to better predictions
•Historical Data: At least 2-3 months of data recommended for reliable predictions
•Data Consistency: Regular attendance tracking and progress updates improve accuracy
•External Factors: AI cannot account for external factors (family issues, injuries, etc.)

Limitations of AI Predictions

•Not Deterministic: Predictions are probabilities, not certainties
•Limited Context: AI doesn't know about personal circumstances or recent conversations
•Requires Human Judgment: Use AI insights as one tool among many, not the sole decision-maker
•New Students: Predictions are less reliable for students with limited historical data

Important: AI as a Tool, Not a Replacement

AI insights are designed to augment your expertise as an instructor, not replace it. Always combine AI predictions with your personal knowledge of each student, recent conversations, and professional judgment. Trust your instincts and use AI as one of many data points.

10. Common Pitfalls and Troubleshooting

Here are solutions to common issues when using AI-powered insights:

❌ Problem: "AI features are not available" error

Cause: Your club does not have the AI-Enabled subscription tier.

Solution:

1. Navigate to Dashboard → Billing
2. Click "Upgrade to AI-Enabled"
3. Complete the subscription upgrade (A$199/month)
4. Wait 1-2 minutes for subscription to activate
5. Refresh the page and try again

Note: If you recently upgraded and still see this error, ensure your subscription status is "active" in the billing page.

❌ Problem: No insights appear for a student

Cause: Insights have not been generated yet for this student.

Solution:

1. Open the student's profile page
2. Scroll to the AI Insights section
3. Click "Analyze Student with AI"
4. Wait 5-10 seconds for analysis to complete
5. Insights will appear once generated

Alternative: Use "Analyze All Students" in the AI Insights Dashboard to generate insights for all students at once.

❌ Problem: Insights are marked as "stale" or outdated

Cause: Insights are older than 24 hours and may not reflect recent changes.

Solution:

1. Click "Refresh Insights" or "Analyze Student with AI"
2. Wait for new insights to be generated
3. New insights will replace the stale ones

Best Practice: Regenerate insights weekly or after significant events (grading, long absence, etc.).

❌ Problem: Low confidence scores (below 50%)

Cause: Insufficient historical data for the student.

Solution:

1. Ensure attendance is being tracked consistently
2. Record stripe awards and skill progress regularly
3. Wait for 2-3 months of data to accumulate
4. Regenerate insights after more data is available

Note: Low confidence scores are normal for new students. Use predictions cautiously and rely more on direct observation.

❌ Problem: Retention risk prediction seems inaccurate

Cause: AI doesn't know about external factors or recent conversations.

Solution:

1. Review the specific factors contributing to the risk score
2. Consider external factors AI cannot see (family issues, injuries, schedule conflicts)
3. Use AI prediction as one data point, not the only factor
4. Trust your personal knowledge of the student
5. Update student notes with context to help future predictions

Remember: AI predictions are probabilities based on patterns, not certainties. Your expertise and personal relationships are irreplaceable.

❌ Problem: "Analyze All Students" fails or times out

Cause: Large number of students or API rate limiting.

Solution:

1. Try analyzing students in smaller batches (by belt rank or class)
2. Wait 5-10 minutes and try again
3. Check your internet connection
4. If problem persists, contact support

Note: For clubs with 100+ students, bulk analysis may take several minutes. The system processes students in batches to avoid timeouts.

❌ Problem: Grading readiness shows "READY" but student doesn't seem ready

Cause: AI evaluates quantitative metrics (attendance, stripes, training days) but cannot assess qualitative factors (technique quality, attitude, focus).

Solution:

1. Review the "Areas to Improve" section in the prediction
2. Conduct a personal assessment of technique quality
3. Consider the student's mental readiness and confidence
4. Use AI readiness as a starting point, not final decision
5. Delay grading if you have concerns, regardless of AI prediction

Remember: You are the expert. AI provides data-driven insights, but you make the final grading decisions based on holistic assessment.

❌ Problem: AI insights not appearing in Student Portal

Cause: Insights have not been generated or student doesn't have portal access.

Solution:

1. Verify insights have been generated for the student
2. Ensure student has STUDENT role assigned
3. Check that student has login credentials
4. Ask student to log out and log back in
5. Regenerate insights if they're older than 24 hours

Note: Students only see motivational messages and progress trajectory, not retention risk scores or detailed instructor recommendations.

11. Related Features

AI-powered insights work best when combined with other DojoMaster features:

📈 Student Progress Tracking →

Track student development over time with detailed progress metrics that feed into AI predictions.

✅ Attendance Tracking →

Accurate attendance data is crucial for retention risk predictions and grading readiness.

🥋 Grading Event Setup →

Use grading readiness predictions to identify which students to enroll in upcoming grading events.

🎯 Stripe/Theme System →

Stripe completion percentage is a key factor in grading readiness predictions.

👥 Student Enrollment →

Manage student information and enrollment status that AI uses for predictions.

📊 Dashboard Overview →

Access AI insights from the main dashboard with quick links to at-risk students and grading-ready students.

👨‍👩‍👧 Parent Portal →

Parents can view their child's AI insights including progress trajectory and motivational messages.

🔔 Notifications →

Set up automated notifications for high-risk students or students ready for grading.

12. Next Steps

Ready to start using AI-powered insights? Follow these steps:

Verify AI-Enabled Subscription

Navigate to Dashboard → Billing and confirm you have the AI-Enabled tier (A$199/month). If not, click "Upgrade to AI-Enabled" to activate AI features.

Navigate to AI Insights Dashboard

Go to Dashboard → AI Insights to access the AI insights hub. Familiarize yourself with the five tabs: Overview, At-Risk Students, Ready for Grading, Top Performers, and All Insights.

Generate Insights for All Students

Click "Analyze All Students" to generate AI insights for your entire club. This may take a few minutes depending on the number of students. Once complete, you'll see club-wide analytics and individual predictions.

Review At-Risk Students

Switch to the "At-Risk Students" tab to see students with MEDIUM, HIGH, or CRITICAL retention risk. Review the risk factors and recommendations for each student. Prioritize CRITICAL and HIGH risk students for immediate intervention.

Identify Students Ready for Grading

Switch to the "Ready for Grading" tab to see students with READY or HIGHLY_READY grading readiness. Review their target belt, readiness factors, and estimated time to readiness. Consider enrolling them in your next grading event.

Schedule Regular Insight Reviews

Set a recurring calendar reminder to review AI insights weekly or monthly. Regenerate insights regularly (they're cached for 24 hours) to ensure predictions reflect current student status. Use insights to inform your teaching strategy and student engagement efforts.

🎯 Pro Tip: Combine AI with Personal Touch

The most effective use of AI insights is to combine data-driven predictions with your personal knowledge of each student. Use AI to identify students who need attention, then apply your expertise and relationship-building skills to address their specific needs. AI augments your teaching, it doesn't replace it.

← Back to Documentation

Kiosk Mode

Student Progress Tracking

Back to Documentation Hub