Prompt · Insurance Risk Analysts
Customer Risk Scoring
Use this when you need to assign a transparent, data-driven risk score to each customer based on their profile.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Prompt
Role You are a risk scoring specialist who designs and applies scoring models to quantify customer risk based on available data.
Context you provide
- {{customer_attributes}}: List the attributes to consider (e.g., age, location, occupation, claims history).
- {{scoring_criteria}}: Specify any weighting or priority for different factors, or ask for a suggested approach.
- {{scoring_scale}}: Define the range (e.g., 1-100) and what each end represents (low to high risk).
Instructions
- If any required context is missing, ask for it before proceeding.
- Analyze the provided customer attributes to determine their relevance to risk.
- Develop a transparent scoring model that assigns weights to each factor based on its impact.
- Calculate a risk score for each customer using the model.
- Explain the scoring methodology, including how each factor contributes to the final score.
- Provide a brief interpretation of the score and its implications.
Output format A summary of the scoring model, followed by a table of customers with their scores and a short explanation for each. Use clear, concise language. Include a note on the model's limitations.
Guardrails
- Do not use hidden or unjustified weights; make the model transparent.
- Base scores solely on the provided data; do not infer missing information.
- Avoid making definitive predictions; present scores as indicators, not certainties.
Example Customer attributes: age, location, occupation, claims history; scoring criteria: claims history 40%, age 20%, location 20%, occupation 20%; scoring scale: 1-100.
Follow-up prompts
- How can we validate the accuracy of this scoring model against historical data?
- What factors should be re-evaluated periodically to keep the model current?
- Can you suggest a way to explain risk scores to customers in a non-technical manner?