The post-covid "new normal" has broken the models Banks use to understand credit risk: economies, markets, businesses and consumers all behave very differently than they did when the models were built. Banks need to rebuild hundreds of models, but, with each model taking up to 9 months, a much faster and more efficient solution is urgently needed.

However, automation alone is not enough - generic machine learning tools looking for statistical patterns in historical data just speed-up the process of getting the wrong answer.

RISKROBOT™ by SPIN Analytics is a new approach which combines Expert Judgement with Artificial Intelligence (AI) and Machine Learning (ML) techniques on Big Data, to produce predictive analytics which are accurate and relevant for making credit decisions and managing risk. RISKROBOT™ supports all types of credit risk models (decisioning, BIII/BIV RWA, IFRS9/CECL, stress testing) for all types of retail and commercial products and portfolios. It is encoded with all current Credit Risk Regulation and decades of experience so that human experts are augmented - not replaced.

With RISKROBOT™, financial institutions can now:

Rapidly respond to market changed combining data analysis with expert judgement.
Reduce the total time to update or replace models by more than a factor of 10.
Shorten the delivery timelines of credit documentation from months to minutes.
Use credit risk experts more efficiently on analysis and applying judgement by automating routine tasks.

RISKROBOT™ key technological components will help you:

Manage, homogenise and understand your data for efficient processing leading to a high-performance Credit Risk Data Warehouse maximising credit insights.
Keep your model up to date via sound statistical techniques and matching current regulations.
Orchestrate assumptions & analytical techniques across all models so credit risk is consistently calculated and explainable across the portfolio.
Automate the generation of regulatory documents including audit and reports.

spinanalytics-riskrobot-pilot

avtor: SPIN Analytics