Part 1: Introduction & The Adaptive Learning Problem (10 minutes)
0:00 - 0:10
Quick Poll
"What's your biggest challenge in personalizing learning?"
- Too many students, not enough time
- Can't diagnose individual gaps
- Students at different levels
- Keeping advanced students challenged
The Problem (5 minutes)
Three critical challenges:
- Philosophy rigidity: One-size-fits-all pedagogical approach
- AI opacity: Black-box decisions without transparency
- Narrow focus: ITS that only do one thing well
UALS Solution (5 minutes)
Four key innovations:
- Multi-philosophy: Curriculum OR competency-based (same system!)
- Three modes: Exploration (KE), Inquiry (SPL), Assessment (SBCAT)
- 100 agents: Comprehensive AI support (vs. typical 3-5 agents)
- Explainable: "Show AI Thinking" reveals decision-making
View Dashboard
Part 2: Three Learning Modes - Live Demo (25 minutes)
0:10 - 0:35
Knowledge Explorer (KE) - 8 minutes
Theory (2 min):
- Constructivist exploration with concept mapping
- Visual knowledge structures + progressive disclosure
- Adapts complexity (novice → intermediate → advanced)
Live Demo (6 min):
- Navigate to curriculum mode: Biology → Photosynthesis
- Show interactive concept map
- Click concept to see: definition, visual, examples, practice questions
- Demo prerequisite locking/unlocking
- Show how content adapts to mastery level
Key Takeaway: Self-directed exploration with just-in-time support
Try Knowledge Explorer
Socratic Playground (SPL) - 8 minutes
Theory (2 min):
- Guided inquiry through problem-solving
- Five hint agents (conceptual, procedural, reflection, worked example, error analysis)
- Socratic questioning > telling answers
Live Demo (6 min):
- Launch SPL from photosynthesis topic
- Show authentic scenario (plant health problem)
- Demonstrate hint progression:
- Try problem without hints (productive struggle)
- Request conceptual hint (guiding question)
- Request procedural hint (next step)
- Show worked example if needed
- Demo misconception correction through EMT dialogue
Key Takeaway: Learning through scaffolded problem-solving
Try Socratic Playground
Scenario-Based CAT (SBCAT) - 9 minutes
Theory (2 min):
- Adaptive assessment using Item Response Theory
- Difficulty adjusts to learner (θ estimation)
- Diagnostic feedback + formative guidance
Live Demo (7 min):
- Launch SBCAT assessment
- Show adaptive difficulty:
- Answer easy question correctly → harder question
- Answer hard question incorrectly → easier question
- Demonstrate feedback (not just "wrong"—why and how to improve)
- Show final diagnostic report:
- Ability estimate (θ = 0.75, intermediate)
- Strengths identified
- Gaps for remediation
- Recommended next steps (KE topics, SPL problems)
Key Takeaway: Assessment as learning opportunity, not just measurement
Try Adaptive Assessment
Part 3: The 100-Agent "Show AI Thinking" Feature (15 minutes)
0:35 - 0:50
Agent Architecture Overview (5 min)
100 specialized agents across 17 categories:
- Learner modeling, pedagogical strategies, content generation
- Assessment, cognitive support, social-emotional learning
- AI ethics, neuroscience, domain specialists, longitudinal learning
10 coordination levels: Lightning (4 agents, 1s) → Complete (100 agents, 40s)
Production-ready: 30 core agents + 70 extended (4-year roadmap)
"Show AI Thinking" Demo (10 min)
Setup:
- Click "Get AI Recommendation" from dashboard
- Enter goal: "Understand photosynthesis"
- Select Analysis Level 2 (Quick, 11 agents, 2-4s)
- Submit
Results Screen:
- System recommends: "Socratic Playground (SPL)"
- Rationale: "You have foundational knowledge but need application practice"
Click "🧠 Show AI Thinking" button:
- Visual workflow timeline appears
- Shows 11 agent invocations in sequence
- Playback controls: play/pause, speed adjustment (0.5x, 1x, 2x, 4x)
- Click individual agents to see detailed reasoning
Key Takeaway: Complete transparency builds trust in AI recommendations
Try AI Recommendation
Part 4: Use Cases & Next Steps (10 minutes)
0:50 - 1:00
Quick Use Cases (3 min)
Classroom Supplementation:
- KE for homework → SPL for in-class practice → SBCAT for exit tickets
Online Course:
- Self-paced with AI recommendations guiding KE → SPL → SBCAT cycles
Remediation:
- SBCAT pre-assessment identifies gaps → KE fills gaps → SPL builds skills
Corporate Training:
- Competency-based onboarding → SPL job scenarios → SBCAT certification
Q&A (5 min)
Anticipated Questions:
- "How much does it cost?" → LLM costs <$0.01/student/hour (94% cache hit rate)
- "Can I customize content?" → Yes, custom prompts + domain model import
- "What about incorrect LLM content?" → Validation layers + educator review dashboard
- "How long to implement?" → Pilot in 2-4 weeks, full deployment 2-3 months
Next Steps (2 min)
Takeaways:
- Workshop slides: Sent via email today
- Demo access: 30-day trial account credentials
- Documentation: Links to guides, architecture, research articles
Call to Action:
"Interested in a pilot program? Let's schedule a 30-minute follow-up call to discuss your specific needs."
Contact:
- Email: pilot@uals.edu
- Office hours: First Friday of each month, 10 AM PT
- Community: Join Discord/Slack for UALS educators