🔒 Answer Keys Conceptual

Password-Protected Instructor Material

Conceptual Answer Keys - Exercises 1 & 2

Purpose: Show instructors what “good thinking” looks like (not prescriptive answers)

Exercise 1: AI Tech Radar - Conceptual Answers

What “Good” Looks Like

Not: All cards in correct quadrants (there are no “correct” quadrants)

But: Teams that: 1. Discuss WHY each card belongs where 2. Notice patterns in their placement 3. Question whether patterns are intentional 4. Recognize clustering in Optimize/H1 5. Ask: “Should we be more balanced?”

Example Good Thinking (Not “Right Answer”)

Team discussion you want to hear: - “Most of our cards are Optimize, Horizon 1” - “Why is Transform empty?” - “Transformation feels risky without data readiness” - “But if we ONLY optimize, what happens when competitors transform?” - “So maybe we ADD one Transform initiative?”

Notice: No single “right” answer. The thinking matters more than placement.

Common Placement Patterns (All OK)

Pattern 1: Heavy Optimize, Light Transform - Why teams do this: “It’s safer, has better ROI” - What matters: “Are you aware of this choice and its risks?” - Good outcome: Team says “yes, intentional” or “wait, maybe we should change”

Pattern 2: Most H1/H2, Few H3 - Why: “H3 is too uncertain” - What matters: “Is that strategic or just default?” - Good outcome: Team debates whether competitive pressures force more H3 investment

Pattern 3: Even spread across quadrants - Why: Unusual but possible (some teams strategically balance) - What matters: “How did you think about balance?” - Good outcome: Team articulates portfolio strategy

Indicators of Weak Thinking

❌ Cards placed without discussion (“just our gut”)
❌ Dismissing Transform entirely (“too risky”)
❌ Not noticing clustering pattern
❌ Defensive when you point out patterns (“That’s just how we operate”)

Facilitation Moves During Debrief

If team has heavy Optimize clustering: You: “I’m noticing 8 of 10 cards here. Is that intentional or default?” Team: “Just how we naturally think” You: “Got it. Question: If competitors invest 30% in Transform, what happens to you if you stay 90% Optimize?”

Don’t tell them to change. Let them realise the strategic risk.

If team dismisses Transform: You: “What would a transformational AI initiative look like for you?” Team: “I don’t know, it feels too uncertain” You: “That uncertainty—is it ‘impossible’ or ‘we haven’t thought about it’?”

Opens space for exploration.

What Teams Learn (Not Outputs)

✓ Strategic balance isn’t accidental ✓ Most organisations cluster in Optimize by default ✓ Intentional portfolio strategy requires explicit conversation ✓ Transform initiatives feel risky because they ARE—but that’s the point

Exercise 2: AI-Assisted Strategic Analysis - Conceptual Answers

What “Good” Engagement Looks Like

Not: Accepting AI output as truth

But: Teams that: 1. Read AI responses carefully 2. Challenge or build on AI’s suggestions 3. Extract useful insights even from “meh” responses 4. Combine human expertise with AI breadth 5. Ask follow-up questions to deepen

Example Good Thinking (Not “Correct Responses”)

Round 1: Risk Deep-Dive

Weak engagement: Team reads AI response: “Yeah, that’s generic. Our industry is different.” Dismisses, moves on.

Good engagement: Team reads AI response: “AI says competitors will use embedded finance. That’s not our immediate threat, but… wait, what if app-based retailers START embedding our competitors’ products?” Extracts kernel of insight, builds on it with their context.

Key difference: Not whether AI was “right,” but whether team extracted value.

Sample Round 1 Insight Extraction (Retail Context)

What AI typically generates: - Platform aggregators (Google Shopping, Pinterest) bypassing your website - Suppliers going direct-to-consumer with AI - Gen Z expectations about AI personalisation - Second-order effects of dynamic pricing (data flywheel) - Customer data devaluation as competitors learn faster

What good teams do with this: - Pick ONE insight: “The Gen Z expectations gap is real for us” - Connect to their business: “We’re averaging customer age 38. Gen Z is already shopping elsewhere.” - Build on it: “So our priority should be personalisation, not pricing optimization?” - Decide to act: “Add to our H2 roadmap”

You observe: Team didn’t accept AI wholesale. They evaluated, connected to their context, decided what matters.

Sample Round 2: Reverse Prompting (Common Pattern)

What typically happens:

AI Question 1: “What manual process takes the most human time?” Team answer: “Documentation—nurses spend 3 hours/shift on clinical notes”

AI Question 2: “What percentage is duplicative information already captured elsewhere?” Team answer: “Maybe 50%? We enter vitals three places…”

AI Question 3: “If AI auto-populated duplicates, how would nurses use freed time?” Team answer: “Oh! More patient interaction. That improves satisfaction AND outcomes.”

AI Question 4: “What prevents you from implementing this today?” Team answer: “Data integration—systems don’t talk. Regulatory concerns.”

The “aha!” moment: Team went from “we’ll never automate clinical work” to “actually, admin documentation automation could be a gateway.”

What you observe: Reverse prompting revealed opportunity hiding in complaint.

Sample Round 3: AI Debate Pattern

Expected outcome: Teams see both sides have merit - Optimist: “Speed matters. Competitors are moving fast.” - Pragmatist: “Data foundation matters. Bad data kills projects.” - Synthesis: “Do both—pilot on best data while building infrastructure. But discipline on scaling gates.”

Good team response: “We’ve been arguing about this internally. Seeing both sides articulated makes it clear: we need to do both, not choose.”

Indicators of Weak Engagement

❌ Dismissing all AI output (“This doesn’t apply to us”)
❌ Accepting all AI output uncritically (“AI said it, so we’ll do it”)
❌ Shallow prompts (no industry context = generic responses)
❌ Not discussing what AI said (silently reading)

Indicators of Strong Engagement

✅ Team discusses AI response aloud
✅ Someone asks follow-up prompt
✅ Debate about whether AI’s suggestion applies to them
✅ Capturing AI insights that change their thinking
✅ “Wait, what if we combined AI’s point with what we already know?”

Facilitation Moves During Phase 3 Synthesis

If table seems unengaged: You: “What’s ONE thing AI said that you hadn’t considered?” Team: “Well… the second-order effect about data…” You: “What makes that useful?”

Helps them extract value from “meh” responses.

If table accepted AI output blindly: You: “Is that actually true for YOUR business or is AI generalizing?” Team: “Oh, that’s a good question. We’d need to…” You: “Exactly. AI generates, you evaluate. Both parts matter.”

If table is debating AI accuracy: You: “You’re right, AI isn’t always accurate. But did it change how you THINK about the problem?” Team: “Yeah, actually…” You: “That’s the value. Not accuracy, but different perspective.”

What Teams Learn (Not Outputs)

✓ AI as thought partner (generates ideas, questions, perspectives) ✓ Better prompts = better AI insights (quality in, quality out) ✓ Humans bring context, AI brings breadth ✓ AI doesn’t replace judgment—it augments it ✓ Some AI responses suck, but even weak AI often has one useful idea

What NOT to Grade As “Correct”

Exercise 1

❌ Specific quadrant placements (variation is healthy)
❌ Number of Transform cards (some balance is good, not mandated)
❌ Specific H1/H2/H3 distribution (depends on strategy)

Exercise 2

❌ Whether they used specific AI tools (Gemini vs. ChatGPT doesn’t matter)
❌ Whether AI responses were “high quality” (even medium responses teach)
❌ Whether they adopted AI suggestions (they might validly reject them)

What to Assess AS “Good Thinking”

Exercise 1

✅ Team discusses placement (not silent) ✅ Notices patterns in their own choices ✅ Questions whether patterns are strategic vs. accidental ✅ At least considers what balance would mean

Exercise 2

✅ Engages with AI output (reads, discusses, questions) ✅ Extracts insights even from mediocre responses ✅ Asks follow-up prompts ✅ Combines human expertise + AI breadth ✅ Evaluates rather than accepts/rejects wholesale

Calibration During Debrief

When a team shares their learning:

You hear: “AI identified competitive threats from adjacent industries” You think: Good—they extracted breadth value from AI

You hear: “We realised we’ve been optimizing when we should transform” You think: Good—frameworks + exercise created insight

You hear: “AI said X and we’re going to do it” You think: Maybe—check if they evaluated or just accepted

You hear: “We rejected AI’s suggestion because it doesn’t fit our business” You think: Good—they evaluated critically

The goal isn’t right answers. It’s teaching teams to THINK about AI differently: as augmentation, not replacement or oracle.