Chain of Thought Prompt Examples

A chain-of-thought prompt that walks through a structured decision framework, considering multiple factors, trade-offs, and scenarios before arriving at a recommendation.

Analyse this business decision step by step. Show your reasoning at each stage.

Decision: [DESCRIBE THE DECISION TO BE MADE]
Options: [LIST THE OPTIONS]
Context: [RELEVANT BUSINESS CONTEXT]

Work through this framework:

**Step 1: Frame the decision**
- What exactly are we deciding?
- What are the constraints?
- What would "success" look like?

**Step 2: Gather and assess the facts**
- What do we know for certain?
- What are we assuming? (Flag each assumption explicitly)
- What don't we know that would change the decision?

**Step 3: Evaluate each option**
For each option, reason through:
- Expected outcome in the best case
- Expected outcome in the worst case
- Probability-weighted expected value
- Key risks and how they could be mitigated

**Step 4: Consider second-order effects**
- What happens 6-12 months after this decision?
- How will competitors, customers, or employees react?
- What doors does this open or close?

**Step 5: Make a recommendation**
- State your recommendation clearly
- Explain the primary reason it's the best choice
- List the conditions under which you'd change your recommendation
- Suggest a "kill criteria" — when should we reverse course?

Be explicit about your reasoning at every step. If you're uncertain, say so and explain why.

A structured troubleshooting prompt that systematically narrows down root causes through hypothesis generation, testing, and elimination.

Troubleshoot this technical issue step by step.

Problem: [DESCRIBE THE ISSUE]
System: [WHAT SYSTEM/APPLICATION]
When it started: [TIMESTAMP OR CONTEXT]
What changed recently: [ANY RECENT CHANGES]
Error messages: [PASTE ERRORS IF ANY]

Work through systematically:

**Step 1: Understand the symptoms**
- What exactly is failing?
- What is still working?
- Is it consistent or intermittent?
- Who/what is affected?

**Step 2: Generate hypotheses**
List the most likely causes in order of probability:
1. [Most likely cause] — because [reasoning]
2. [Second most likely] — because [reasoning]
3. [Third most likely] — because [reasoning]

**Step 3: Diagnostic plan**
For each hypothesis, what would confirm or rule it out?
- Hypothesis 1: Check [X]. If [result], then this is/isn't the cause.
- Hypothesis 2: Check [Y]. If [result], then this is/isn't the cause.

**Step 4: Narrow down**
Based on available information, which hypotheses can we eliminate?
Show your reasoning for each elimination.

**Step 5: Root cause and fix**
- Most likely root cause: [X]
- Confidence level: [high/medium/low]
- Recommended fix: [specific steps]
- How to verify the fix worked: [verification steps]
- How to prevent recurrence: [preventive measures]

If you need more information to proceed at any step, say exactly what information you need and why.

Walks through a systematic risk assessment for a project or initiative, identifying risks, evaluating probability and impact, and developing mitigation strategies.

Conduct a risk assessment for: [PROJECT/INITIATIVE]

Context: [DESCRIBE THE PROJECT AND ITS OBJECTIVES]
Timeline: [DURATION]
Stakeholders: [KEY STAKEHOLDERS]

Work through each step:

**Step 1: Identify risks**
Think through risks in each category:
- Technical risks: What could go wrong technically?
- Resource risks: People, budget, time constraints
- External risks: Market, regulatory, vendor dependencies
- Operational risks: Process, change management, adoption
List each risk with a brief description.

**Step 2: Assess each risk**
For each risk identified, reason through:
- Likelihood (1-5): Why do you rate it this way? What evidence supports this rating?
- Impact (1-5): If this happens, what would the consequences be? On timeline? On budget? On quality?
- Risk score: Likelihood × Impact
- Show your reasoning for each rating

**Step 3: Prioritise**
Rank risks by score. For the top 5:
- What are the early warning signs?
- At what point would this risk become critical?

**Step 4: Mitigation strategies**
For each top-5 risk:
- Prevention: What can we do to reduce likelihood?
- Contingency: If it happens, what's our response plan?
- Owner: Who should be responsible for monitoring and responding?
- Cost of mitigation vs cost of risk materialising

**Step 5: Residual risk assessment**
After mitigations, what is the residual risk level? Is it acceptable?

Present the final risk register as a table with columns: Risk, Likelihood, Impact, Score, Mitigation, Owner, Status.

Guides AI through interpreting complex data by working through each observation, checking for alternative explanations, and building to a conclusion.

Interpret this data step by step. Do not jump to conclusions.

Data:
[PASTE YOUR DATA — tables, metrics, observations]

Context: [WHAT THIS DATA REPRESENTS AND WHY WE'RE LOOKING AT IT]

Work through:

**Step 1: Observe**
- What patterns do you see in the data?
- What stands out as unusual?
- What's missing that you'd expect to see?
List each observation as a factual statement, no interpretation yet.

**Step 2: Generate explanations**
For each notable observation, generate at least 2 possible explanations:
- Observation: [X]
  - Explanation A: [reason] — how well does the data support this?
  - Explanation B: [reason] — how well does the data support this?

**Step 3: Test explanations**
Which explanations are consistent with ALL the data, not just the single observation?
Eliminate explanations that contradict other data points. Show your reasoning.

**Step 4: Check for confounding factors**
- Could external factors explain the patterns? (seasonality, market events, data quality issues)
- Are there lurking variables that might create a spurious correlation?
- Is the sample size sufficient for the conclusions we're drawing?

**Step 5: Conclusions**
- State your main findings (supported by the data)
- Rate your confidence: High / Medium / Low for each finding
- What additional data would increase your confidence?

**Step 6: So what?**
- What actions should we take based on these findings?
- What should we investigate further?

Applies a structured ethical reasoning framework to a business or technology decision, considering multiple ethical perspectives and stakeholder impacts.

Analyse the ethical dimensions of this decision step by step.

Decision: [DESCRIBE THE DECISION]
Context: [RELEVANT BACKGROUND]
Stakeholders: [WHO IS AFFECTED]

**Step 1: Identify the ethical dimensions**
- What are the ethical questions embedded in this decision?
- Who could be helped or harmed?
- What rights or values are at stake?

**Step 2: Apply ethical frameworks**

Utilitarian perspective: What outcome produces the greatest good for the greatest number?
- Who benefits and how?
- Who is harmed and how?
- Net outcome assessment

Deontological perspective: What are our duties and obligations regardless of outcome?
- What rights must be respected?
- What rules or principles apply?
- Would we be comfortable if this decision became universal policy?

Virtue ethics perspective: What would a person of good character do?
- Does this decision align with values like honesty, fairness, compassion?
- Would we be proud to explain this decision publicly?

**Step 3: Identify tensions**
Where do the frameworks disagree? Which tensions must be resolved?

**Step 4: Stakeholder impact**
For each stakeholder group, what is the impact and have they had a voice in the decision?

**Step 5: Recommendation**
- Recommended course of action
- What safeguards or conditions should be attached?
- How to communicate the decision transparently
- What to monitor for unintended consequences

Uses game theory and strategic reasoning to think through competitive moves, anticipate responses, and identify optimal strategies.

Reason through this competitive strategy question step by step.

Situation: [DESCRIBE THE COMPETITIVE SITUATION]
Our position: [OUR CURRENT MARKET POSITION AND RESOURCES]
Key competitors: [LIST MAIN COMPETITORS AND THEIR POSITIONS]
Decision: [WHAT STRATEGIC MOVE ARE WE CONSIDERING?]

**Step 1: Current state analysis**
- What is the competitive equilibrium right now?
- What advantages and disadvantages does each player have?
- What are each competitor's likely priorities?

**Step 2: First-order effects**
If we take this action:
- How does it change our position?
- What message does it send to the market?
- What are the immediate costs and benefits?

**Step 3: Competitor response prediction**
For each key competitor:
- What is their most likely response? Why?
- What is their most dangerous possible response?
- What constraints limit their response options?

**Step 4: Second-order effects**
After competitors respond:
- Where does the market equilibrium settle?
- Are we better or worse off than before?
- What options does this open or close for us?

**Step 5: Alternative strategies**
What other moves could we make instead?
For each alternative, briefly assess first and second-order effects.

**Step 6: Recommendation**
- Recommended strategy and sequence of moves
- Key assumptions that must be true for this to work
- Contingency plan if competitors respond differently than expected
- Metrics to track to know if the strategy is working