AI Innovation Challenge · Week 3 of 4

This is Your
Dress Rehearsal.

Next week is the real thing. Today, every team pitches. Every team gets feedback. And every team leaves better than they arrived.

🎤

Full Pitch

Every team presents their complete deck today. Time it.

🔍

Peer Feedback

Structured critique from classmates using the rubric.

✏️

Refine

Leave with specific improvements. Don't leave with "it was fine."

Today's Session · 60 min

What We're Doing Today

0–5 minRubric walkthroughEveryone understands how judging works before pitching
5–10 minDelivery tips — 5 quick principlesBody language, pacing, handling questions
10–45 minPractice pitches — all teams5 min per team + 3 min structured peer feedback
45–55 minTeacher feedback highlightsPatterns you noticed across teams — what needs fixing
55–60 minExhibition logistics + final checklistWhat to expect next week, who votes, how it works
Teacher: With ~5 teams per section and 8 min each, this fits tight. Keep transitions snappy. Use a visible timer.
Judging Criteria

The Rubric

Criteria3 — Strong2 — Developing1 — Weak
Problem ClarityReal, specific, compellingSomewhat clearVague or obvious
AI Solution + ML TypeNamed, justified, explained clearlyNamed but vagueMissing or wrong
Data & BiasSources + type + mitigationPartial — missing mitigationNot addressed
Ethics & ImpactSpecific harm + safeguards namedAcknowledged, not addressedNot addressed
Limitations & Failures2+ specific failure casesVague acknowledgment"It will always work"
Future WorkAmbitious, specific, groundedGeneric ("add features")Not included
Pitch QualityClear, confident, good visualsMostly clear, some rough spotsHard to follow
Teacher: Print this for peer feedback. 50% judge score, 50% audience vote at the exhibition.
Pitch Delivery

5 Things That Separate
Good Pitches from Great Ones

🎯 Lead with the problem

  • Don't open with "Hi, we're Team X and today we're going to talk about..."
  • Open with a story, stat, or scenario that makes the audience feel the problem.

👁️ Slides support you

  • You are the pitch — slides are visual aids, not scripts
  • No slide should be a wall of text. If you wrote it to read aloud, cut it.

⏱️ Respect the clock

  • 5–7 minutes. Practice out loud. Time yourself.
  • Going over signals you haven't prepared.

🧠 Know your content

  • Every team member should be able to answer questions — not just the speaker
  • Judges will ask the quiet ones.

💬 Handle questions well

  • "That's a great point, and here's how we've thought about it..." is a strong opener
  • "We haven't thought about that" is honest — but have a plan B.

✅ Own your limitations

  • Pretending your system is perfect loses judge trust immediately
  • Honest acknowledgment of bias and failure builds credibility.
Teacher: Do a quick demo — give a 60-second bad pitch vs a 60-second good pitch on the same idea. Makes the principles stick.
Peer Feedback

How to Give Useful Feedback

After each team pitches, the class gives structured feedback. Here's how:

🚫 What NOT to say in feedback
  • "It was really good / great / amazing" — what specifically worked?
  • "I didn't understand anything" — what part specifically?
  • "Your idea is bad" — challenge the design, not the concept
Teacher: Model the feedback yourself after the first team. Show them what "specific and useful" looks like.
Watch Out For

Most Common Mistakes
in Student AI Pitches

❌ "Our AI will be 99% accurate"

Overconfident claims with no basis. Judges don't trust teams who haven't thought about failure. Accuracy claims need context, benchmarks, and acknowledgment of edge cases.

❌ "We'll just use good data"

Not a bias mitigation strategy. Judges want to know which bias types apply, what specific steps address them, and what monitoring looks like post-deployment.

❌ "Ethics: our system is safe"

Empty claim. Every AI system has ethical implications. "Safe" needs to be shown — with safeguards, oversight mechanisms, and acknowledged risks.

❌ Reading off slides word-for-word

If you're reading your slides, the audience could have just read them alone. Know your content. Use slides as cues, not scripts.

Teacher: After practice pitches, call out 1-2 of these patterns you observed. Don't name teams — make it general.
Next Week

What to Expect at the
Exhibition

🎤

The Pitch

5–7 minutes. All four sections participate. Judges + public audience. Be ready for questions after your pitch.

🗳️

The Vote

50% judge rubric scores, 50% audience popular vote. Judges are teachers (and any guests invited). Audience is everyone else.

🏆

The Prize

Top 3 teams announced. The winning idea gets built into a real, working AI system by your teacher. 2nd and 3rd get recognition.

📋

What to Bring

Final slide deck (shared with teacher the night before). Any printed materials or mockups you want displayed. All team members present.

Teacher: Share the exhibition schedule with teams now. Let them know the exact timing so no one is caught off-guard.
Before Exhibition Day

Final Checklist

Content ✓

All 6 pitch requirements covered. Specific on data sources. Specific on bias type + mitigation. Real failure cases. Honest limitations.

Slides ✓

Visuals on every slide. No slide is just text. Diagrams, mockups, or user flows included. Font readable from 3 metres away.

Delivery ✓

Timed at 5–7 min. Everyone knows their section. Practiced out loud at least 3 times. Ready for Q&A on any slide.

You've done the hard work. Now make it shine.

Teacher: Collect final decks the night before. Quick scan to catch any last-minute issues before exhibition day.