Strategic Prompting Guide for AI-Assisted Decision Making

Take-home reference for using AI tools in strategic analysis

---

The Three Techniques You Practiced

1. Risk Deep-Dive

Ask AI to identify what you’re missing or underestimating

2. Reverse Prompting

Have AI ask YOU questions to reveal hidden opportunities

3. AI Debate

Use multiple AIs (or roles) to explore different perspectives

---

1. Risk Deep-Dive Prompt Template

Use when: Evaluating an AI investment, strategic decision, or competitive threat

Copy and customize:

You are a strategic [YOUR INDUSTRY] consultant specializing in AI.

Our team is considering: [DESCRIBE THE INITIATIVE/DECISION]

We've identified these risks:
- [RISK 1]
- [RISK 2]
- [RISK 3]

Help us think deeper:
1. What risks are we missing or underestimating?
2. What second-order effects should we consider?
3. Which risk is most likely to be existential?
4. What early warning signs should we watch for?

Be specific to our industry. Challenge our assumptions.
Use concrete examples from similar situations if relevant.

Follow-up prompts: - “Which risk should we address FIRST and why?” - “What would a mitigation plan for [SPECIFIC RISK] look like?” - “If this initiative fails, what’s the most likely cause?”

What this is good for: - ✅ Identifying blind spots in risk analysis - ✅ Connecting dots between related risks - ✅ Bringing perspectives from adjacent industries - ✅ Challenging overly optimistic assumptions

What this is NOT good for: - ❌ Knowing your specific company’s internal politics - ❌ Understanding proprietary competitive intelligence - ❌ Making the final risk decision (that’s your job)

---

2. Reverse Prompting Template

Use when: Looking for opportunities, trying to break out of conventional thinking

The key insight: Instead of asking AI for answers, ask AI to ask YOU questions

Copy and customize:

You are an expert on [DOMAIN: e.g., AI business innovation, digital transformation].

Your role is to ask ME probing questions to help me discover opportunities 
I'm overlooking.

About my situation:
- Industry: [YOUR INDUSTRY]
- Company type: [YOUR ORGANIZATION]
- Current challenge: [DESCRIBE THE PROBLEM/DECISION]
- Current approach: [WHAT YOU'RE CURRENTLY CONSIDERING]

Your task: Ask me 5 probing questions that will reveal:
- Hidden opportunities I'm not seeing
- Assumptions I should challenge
- Alternative approaches I should consider

IMPORTANT: Ask ONE question at a time. Wait for my answer before 
asking the next question.

Start with your first question now.

How to respond to AI’s questions: - Be honest and specific in your answers - If a question confuses you, that’s often where insight hides—ask for clarification - Don’t rush—take 1-2 minutes to discuss each answer with your team

Example dialogue:

AI: “What manual process takes your team the most time that you’ve never thought to automate?”

You: “Probably the weekly pricing review meeting—4 hours every Monday with 8 people”

AI: “What prevents you from making those pricing decisions more frequently or in real-time?”

You: “Hmm… we’ve always done it weekly. I guess we could do it daily, or… wait, AI could optimize prices continuously…”

AI: “What data would that real-time pricing system need that you don’t currently capture?”

You: “Oh, we’d need competitor pricing data, which we don’t systematically collect. That’s the blocker!”

The power: AI’s questions often reveal opportunities hiding in plain sight.

Advanced technique: After the 5 questions, ask: “Based on my answers, what’s the ONE opportunity I should prioritize and why?”

---

3. AI Debate Template

Use when: Facing a tough decision with valid arguments on both sides

The setup: Give two AIs (or one AI with two roles) opposing positions, let them debate

Method A: Two Different AIs (ChatGPT vs. Claude)

Prompt for AI #1 (e.g., ChatGPT):

Context: We are considering [DESCRIBE THE DECISION/INVESTMENT].

Current situation:
- [KEY FACT 1]
- [KEY FACT 2]
- [KEY CONSTRAINT]

Your role: You are "The Advocate"

Make the STRONGEST case for [OPTION A: e.g., investing now, proceeding aggressively].

Address:
- Why this is strategically necessary
- What we lose if we wait
- How to mitigate the risks

Keep responses to 3-4 sentences. Be persuasive but acknowledge real concerns.

Present your opening argument now.

Prompt for AI #2 (e.g., Claude):

[SAME CONTEXT AS ABOVE]

Your role: You are "The Skeptic"

Make the STRONGEST case for [OPTION B: e.g., waiting, proceeding cautiously].

Address:
- What risks we're underestimating
- Why timing matters
- What we should do BEFORE committing

Keep responses to 3-4 sentences. Be persuasive but acknowledge opportunity costs.

Present your opening argument now.

How to facilitate the debate: 1. Get opening arguments from both AIs 2. Copy Advocate’s argument → Paste into Skeptic’s chat: “Respond to this argument:” 3. Copy Skeptic’s response → Paste into Advocate’s chat: “Counter this point:” 4. Let them debate 2-3 rounds

Method B: One AI, Two Roles (if you only have access to one)

You are facilitating a debate about [THE DECISION].

Participants:
- "The Advocate" argues for [OPTION A]
- "The Skeptic" argues for [OPTION B]

Context: [PROVIDE KEY FACTS]

Please conduct a debate:
1. Advocate presents opening argument (3-4 sentences)
2. Skeptic responds with counterarguments (3-4 sentences)
3. Advocate counters (3-4 sentences)
4. Skeptic makes final point (3-4 sentences)

Label each speaker. Keep each response to 3-4 sentences. Be substantive.

Begin the debate now.

What to look for in the debate: - Which arguments resonate with your gut instinct? - Are there considerations you hadn’t thought of? - Does one side have stronger evidence? - What’s the synthesis position? (Often: “Do X in short term while preparing for Y”)

After the debate, ask AI:

Based on this debate, if you were the decision-maker, what would you decide and why?
What additional information would you need to be confident in the decision?

---

Principles of Effective Prompting

1. Provide Context

Bad: “What AI should we invest in?”
Good: “We’re a $500M retail company, Level 3 data maturity, facing aggressive online competitors. We have $2M to invest. What AI investments should we prioritize?”

2. Be Specific About Output

Bad: “Tell me about AI risks.”
Good: “List the top 5 AI-specific risks for a customer-facing fraud detection system, ranked by likelihood and impact.”

3. Assign a Role

Bad: “What do you think?”
Good: “You are a CFO evaluating this proposal. What financial concerns would you raise?”

4. Request Structured Output

Bad: “Analyse this.”
Good: “Analyse this in 3 parts: (1) Strengths, (2) Weaknesses, (3) Recommended decision with rationale.”

5. Ask AI to Challenge You

Bad: “Is this a good idea?”
Good: “I think this is a good idea. Play devil’s advocate and identify the strongest arguments AGAINST it.”

6. Use Follow-Up Prompts

Don’t accept first response—go deeper: - “Elaborate on point #3” - “What evidence would change your assessment?” - “Give me a concrete example of that risk” - “What would [PERSON/COMPANY] say about this?”

---

Common Use Cases

Strategic Planning

Prompt: “You’re a strategy consultant. We’re developing our 3-year AI roadmap. Ask me 10 questions to ensure we’re thinking comprehensively about threats, opportunities, and organisational readiness.”

Investment Evaluation

Prompt: “We’re evaluating this AI investment: [DETAILS]. Use the AI Investment Checklist framework: assess data readiness (0-10), continuous learning needs, accuracy requirements, explainability, and ethical risks. Be critical.”

Competitive Analysis

Prompt: “You’re our competitor [COMPETITOR NAME] who just hired a new AI-focused CEO. Outline your AI strategy to disrupt our business over the next 18 months. Be aggressive and specific.”

Risk Mitigation

Prompt: “We’re launching this AI system: [DETAILS]. It fails catastrophically 6 months after launch. Walk me through the most likely failure scenarios step-by-step, from initial warning signs to complete failure.”

Change Management

Prompt: “You’re a skeptical VP of Operations who doesn’t trust AI. I’m pitching this AI initiative: [DETAILS]. Push back on it. What are your strongest objections?”

---

What AI Is Good At (Use It For This)

✅ Generating alternatives - “What are 10 different ways to solve this problem?”
✅ Challenging assumptions - “What am I taking for granted that might be wrong?”
✅ Breadth of knowledge - Bringing examples from other industries/contexts
✅ Structured analysis - Breaking complex decisions into components
✅ Playing roles - Devil’s advocate, different stakeholders, competitors
✅ Connecting dots - “If we do X, what second-order effects happen?”
✅ Identifying patterns - “What are common failure modes for this type of initiative?”

---

What AI Is NOT Good At (Don’t Rely On It For This)

❌ Your specific context - AI doesn’t know your company’s politics, culture, history
❌ Proprietary information - AI doesn’t have your competitive intelligence
❌ Gut instinct - AI can’t replicate your 20 years of industry experience
❌ Making the decision - AI provides input, YOU decide
❌ Accountability - If it goes wrong, you can’t blame the AI
❌ Real-time data - AI’s knowledge has a cutoff date
❌ Nuanced judgment - Especially political, ethical, cultural trade-offs

---

The Right Mental Model

AI is your strategic thought partner, not your boss.

Think of AI as: - ✅ A smart colleague who challenges your thinking - ✅ A research assistant who brings different perspectives
- ✅ A brainstorming partner who never runs out of ideas - ✅ A devil’s advocate who stress-tests your assumptions

NOT as: - ❌ An oracle that provides THE answer - ❌ A replacement for your judgment and experience - ❌ A decision-maker (you are still accountable) - ❌ A source of truth about your specific situation

Best results: Human expertise + AI analysis

You bring: - Deep context about your organisation - Industry experience and intuition - Understanding of stakeholders and politics - Ability to judge what’s realistic - Accountability for the decision

AI brings: - Breadth of examples from many industries - Ability to generate many alternatives quickly - Structured analytical frameworks - Tireless questioning and challenge - Different perspectives

Together: Better strategic decisions than either alone.

---

Ethical Considerations

When using AI for strategic decisions:

1. Don’t share confidential information - Don’t paste proprietary financials, competitive intelligence, or sensitive data - Use generic descriptions: “We’re a mid-sized manufacturer” not “Our Q3 EBITDA is…”

2. Verify AI outputs - AI can be confidently wrong - Check facts, especially statistics and specific claims - Treat AI suggestions as hypotheses to validate, not conclusions

3. Cite AI when appropriate - If AI insight influences your decision: “AI analysis suggested we consider [X]” - Helps others understand your reasoning - Models good practice for your organisation

4. Don’t outsource judgment - AI provides input, humans make decisions - You are accountable for outcomes, not the AI - Use AI to inform, not to abdicate responsibility

---

Practice Exercise: Try These Prompts

Beginner:

You are a strategic consultant. I lead [YOUR ROLE] at [YOUR ORGANIZATION TYPE].

What are the 5 most important questions I should be asking about AI investment 
in the next 12 months?

For each question, explain why it matters.

Intermediate:

I'm preparing a board presentation on AI strategy.

You are a skeptical board member who worries about:
- Wasting money on AI hype
- Data security and bias risks
- Distracting from core business

What tough questions will you ask me? Give me your 5 hardest questions.

Advanced:

We're choosing between three AI investments:

A) [DESCRIBE OPTION A]
B) [DESCRIBE OPTION B]  
C) [DESCRIBE OPTION C]

Use the AI Investment Checklist framework (data readiness, continuous learning, 
accuracy, explainability, ethics) to compare them.

Create a scoring matrix. Then recommend which to fund and why.
Be critical—identify red flags.

---

Remember From This Morning’s Exercise

What worked: - Providing clear context before asking questions - Using reverse prompting to uncover blind spots
- Having AI debate different perspectives - Following up on interesting points (not accepting first response) - Combining human intuition with AI analysis

What didn’t work: - Vague prompts that lacked context - Accepting AI output without critical evaluation - Expecting AI to know your specific situation - Treating AI as an oracle instead of a thought partner

The key insight: When you used AI well this morning, it added genuine value to your strategic thinking. When you use it well this afternoon (and Monday morning), it will do the same.

---

Quick Reference Card

Three Templates You Practiced:

1. Risk Deep-Dive “You are a [INDUSTRY] consultant. We’re considering [X]. Help identify risks we’re missing: [1] What are we underestimating? [2] Second-order effects? [3] Early warning signs?”

2. Reverse Prompting
“You are an expert on [TOPIC]. Ask ME questions to help me discover opportunities I’m overlooking about [SITUATION]. Ask one question at a time.”

3. AI Debate “You are ‘The Advocate’ arguing for [OPTION A]. You are ‘The Skeptic’ arguing for [OPTION B]. Context: [SITUATION]. Debate this decision (3-4 sentences each).”

Effective Prompting Rules: 1. ✅ Provide context 2. ✅ Be specific about output 3. ✅ Assign a role 4. ✅ Ask AI to challenge you 5. ✅ Use follow-up prompts

Remember: AI augments your judgment, doesn’t replace it.

---

Use these techniques this afternoon in Dragon’s Den—and Monday morning on real decisions.