An Innovative US Investment Management Firm Improved Recommendation Match Rate from 1.5% to 23% Using ML-Driven Risk Profiling Solution

Challenges

The client faced multiple challenges in accurately profiling HNI investors and optimizing wealth management processes:

• Wealth managers needed structured insights to conduct effective client interviews and portfolio assessments.
• Existing risk profiling models lacked precision, often leading to misaligned investment recommendations.
• Difficulty in classifying customers across five key risk categories — Conservative, Moderately Conservative, Balanced, Moderately Aggressive, and Aggressive.
• Low recommendation match rates, impacting client satisfaction and portfolio performance.

Solution

Coforge implemented a comprehensive machine learning–driven risk profiling and recommendation system to enhance investment insights and decision-making:

• Conducted extensive data preparation and null value handling due to the question–answer nature of the input data.
• Deployed multiple models — Decision Trees, Random Forests, and XGBoost — to improve prediction accuracy and risk segmentation.
• Used misclassification cost analysis to select the most reliable model combination for customer categorization.
• Applied iterative retraining and validation using multiple evaluation metrics to refine accuracy over time.
• Integrated the model outputs into the firm’s advisory workflow to assist wealth managers in providing more accurate recommendations.

The Impact

• Improved recommendation match rate from 5% to 23% within the first deployment cycle.
• Deeper understanding of client investment behavior and preferences.
• Enhanced risk profiling accuracy, enabling more aligned and personalized investment advice.
• Streamlined wealth manager workload through automated insights and prioritized client segmentation.