What You'll Do
The SAIL Challenge develops your ability to work with AI while maintaining your own judgment. You will analyze the same problem three times — first alone, then with AI, then as a decision-maker who must own the outcome.
Total time: approximately 90-120 minutes, depending on case complexity.
The power is in the comparison. By thinking first (Phase 1) and then engaging AI (Phase 2), you can see exactly where AI helps, where it falls short, and where your judgment must lead.
What You Do
Read the case carefully. Analyze it using your own knowledge and reasoning. Produce an initial recommendation or analysis.
Why This Matters
This phase establishes your baseline — what you think before AI influences you. Without this baseline, you cannot evaluate what AI actually contributes. Research shows that productive struggle strengthens learning; this is a "desirable difficulty" that enhances retention and transfer.
Deliverable
Phase 1 Analysis (1-2 pages): Your analysis of the case and initial recommendation, written entirely in your own words based on your own thinking.
What You Do
Now engage AI to work on the same case. Use any AI tool(s) you prefer. But your task is not simply to get an answer — it is to critically evaluate what AI provides.
- Ask AI to analyze the case
- Compare AI's analysis to your Phase 1 analysis
- Identify where AI's output differs from yours
- Evaluate AI's reasoning: What did it get right? What did it miss? Where might it be wrong?
- Check for hallucinations, unsupported claims, or shallow analysis
Why This Matters
This phase develops AI Literacy — understanding what AI can and cannot do in your specific context. The comparison between your thinking and AI's thinking is where learning happens. You are not learning from AI; you are learning to evaluate AI.
Deliverable
Phase 2 AI Evaluation Log (1-2 pages):
- Key prompts you used (copy or summarize)
- Where AI's analysis differed from your Phase 1 analysis
- Where AI added genuine value (new insights, better framing, useful information)
- Where AI fell short (errors, gaps, hallucinations, shallow reasoning)
- What surprised you about the comparison
What You Do
Now synthesize everything into a final Judgment Memo. You are the decision-maker. You have consulted AI (like you might consult a junior analyst), but the decision is yours. You must own it.
Your Judgment Memo Must Address:
- What is your final recommendation?
- Where did you change your mind based on AI input? Why?
- Where did you override or reject AI's suggestions? Why?
- What did you learn about your own thinking through this process?
Why This Matters
This phase develops Leadership and metacognition. AI cannot answer "where did you override AI?" — only you can. By articulating your reasoning, you strengthen your capacity for judgment. This is what employers need: not people who can prompt AI, but people who can evaluate AI and own outcomes.
Deliverable
Judgment Memo (1 page): Your final recommendation with clear articulation of where you followed AI, where you diverged, and why. Written as if you are presenting to a senior decision-maker who wants to understand your reasoning.
Submission Checklist
Before submitting, confirm you have:
- ☐ Phase 1 Analysis — Your initial analysis, completed without AI
- ☐ Phase 2 AI Evaluation Log — Documentation of AI interaction and critical evaluation
- ☐ Phase 3 Judgment Memo — Your final recommendation with reasoning
All three phases must be completed in order. Your Phase 1 must be submitted (or timestamped) before you begin Phase 2.
Common Questions
Can I revise my Phase 1 analysis after seeing AI's response?
No. Phase 1 is your baseline. Changing it defeats the purpose of comparison. Your growth shows in Phase 3, where you synthesize both perspectives.
What AI tools can I use in Phase 2?
Any AI tool is permitted: ChatGPT, Claude, Gemini, Copilot, Perplexity, etc. You may use multiple tools. Document which tools you used.
What if AI gives a better answer than mine?
That's fine — and expected sometimes. The point is not to "beat" AI. It's to develop your ability to evaluate when AI is right, when it's wrong, and when the answer requires judgment AI cannot provide.
How long should each deliverable be?
Quality over quantity. Phase 1: 1-2 pages. Phase 2: 1-2 pages. Phase 3: 1 page. Clear, concise reasoning is valued over length.