Customer churn rate was 12% with no systematic prediction.

Implemented a churn prediction model reducing churn to 7% within 6 months.

Sales forecasts deviated by ±15% each quarter.

Deployed a time‑series model improving forecast accuracy to ±5%.

Inventory stockouts occurred weekly.

Created demand‑forecasting algorithms cutting stockouts by 60%.

Credit risk assessments relied on manual scoring.

Automated risk scoring, decreasing loan default rates by 20%.

Marketing campaign ROI was unknown.

Built uplift models, increasing campaign ROI by 35%.