About Wipro
Wipro brings value to DataRobot implementations and projects through our joint solutions across themes like ML Governance, AI CoE, Model Risk Management and Legacy Modernization. In-built capabilities of DataRobot’s platform in combination with Wipro’s accelerators and frameworks, allow customers to democratize and scale AI across enterprise while addressing operational risks effectively.
Model Risk Management Solution for Governance, Risk & Compliance (GRC) – A Wipro-DataRobot Joint Offering
BFSI enterprises, Fintech companies and financial institutions are increasingly relying on mission-critical predictive models to make decisions. With regulatory standards higher than ever, it is critical for enterprises using machine-learning models to consider the challenges of meeting stringent industry standards and ensuring that developed model is trustworthy, in order to avoid penalty.
Our solution enables you to build, test, deploy and manage all your models while ensuring adherence to regulatory compliances and guidelines. Wipro brings the industry best practices in GRC and MRM to create a rigorous framework, which is aided by DataRobot AI Platform to automate the Data to Value journey. It provides end-to-end model lifecycle support and smart automation features that can be deployed on both cloud and on premise.
Enabling Model–GRC with DataRobot Platform
- Automated Documentation : Automated documentation feature reduces time required for regulatory compliance documentation (Federal Reserve SR-11-7, FDIC FIL-22-2017) that can usually take weeks, or even months, down to seconds.
- Automated Data Quality Assessment & Insight Generation: Instantaneously flag Data Quality issues & rapidly generate EDA results
- Model Monitoring & Continuous AI: Continuously monitor models for drift and accuracy and setup automated retraining workflows based on monitoring results or schedules
- Champion – Challenger Framework: Compare and benchmark multiple challenger models in production to ensure accurate results from best performing model
- Humble & Explainable AI: All models built in the platform are fully explainable. “Humility Rules” can be configured at the time of deployment to prevent and reduce erroneous model outputs
- Automatic Model Bias Detection: Automatically detect and mitigate bias in the model
Rebuilding models to withstand the test of time – a success story of a large insurer
Business Challenge:
- 100+ pricing models need to be rebuilt before the FCA deadline
- Inability to modernize pricing contributed to huge loss
- 4 Strategic projects have had to descope AI due to limited capabilities
Solution:
- Allow actuaries and non-data scientists to build reliable models
- Code free data preparation, model building and deployment
- Robust Governance, Guardrails and Guidance build into the process
- Real-time updates to features, guide rails, guidance and ML algorithms
- One click deployment of models with consistent API model interface
- Use an automated Champion/Challenge approach to optimize models
Outcomes:
- Meet FCA compliance deadlines
- Avoid heavy external consulting costs
- Avoid expensive FTE capacity costs
- Deploy prepared models increasing revenue significantly
- Monitor and improve models to avoid data drifts
- Build advanced models to improve pricing performance
- Reduce IT burden and time spent on environment building by 80%
- Build new models to improve pricing performance vs. competitors