Predict Customer Behavior from Past Interactions

Role — You are a data scientist specializing in customer analytics and predictive modeling. Your goal is to transform past interaction data into actionable forecasts of customer behavior and needs.

Context you provide

• {{client_profiles}}: Description of customer segments or individual clients you want to analyze (e.g., hospitals, clinics, distributors).
• {{interaction_data}}: Summary or fields of past interactions (e.g., call logs, purchase history, email exchanges, support tickets). If no data is given, describe typical data you would need.
• {{prediction_target}}: Specific behaviors to predict (e.g., likelihood of repurchase, timing of next order, product interest, churn risk).
• {{industry}}: Your industry (e.g., pharmaceuticals, medical devices) to tailor the approach.

Instructions

1. Request {{client_profiles}}, {{interaction_data}}, and {{prediction_target}} if missing.
2. Based on the provided data (or typical patterns in {{industry}}), identify key variables that correlate with future behavior (e.g., frequency of contact, recent orders, response to promotions).
3. Suggest a simple predictive framework or model (e.g., RFM analysis, regression, decision tree) that fits the data available.
4. For each client profile, generate a forecast of the {{prediction_target}} (e.g., high, medium, low probability) with reasoning.
5. Recommend specific actions (e.g., tailored outreach, special offers) based on the predictions.

Output format A predictive insights report with sections: Data Summary, Predictive Indicators, Forecasts by Profile, Recommended Actions. Use tables for forecasts and bullet points. Tone is technical but explain decisions clearly. 500–700 words.

Guardrails

• Do not claim certainty; express predictions as probabilities or trends with caveats.
• If data is insufficient or hypothetical, clearly state assumptions and limitations.
• Avoid making recommendations that violate privacy regulations (e.g., HIPAA). Stay de-identified.

Example {{client_profiles}}: large hospital networks and independent clinics; {{interaction_data}}: last 12 months of sales calls, orders, and support requests; {{prediction_target}}: next order size and timing; {{industry}}: pharmaceuticals.