Case Studies

Issue - State of the art policyholder behavior assumptions developed.  And fed into the IFRS17 valuation models.  And causing hard to explain volatility in results quarter after quarter due to minor PHB assumption changes.

​

Solution - I understand both IFRS17 and Predictive Modeling and I can identify and solve the disconnect.  More stable and accurate financial reporting, in line with the retaining effort from the marketing team. 

Issue - AI and Ethic is high on numerous regulators' agenda.  Do you have all the evidences of your anti-discrimination effort, in order to put your best foot forward?  How the NAIC Model Bulletin should be understood?  

​

Solution - My For Humanity Foundations of Independent Audit of AI Systems Certificate acknowledge the expertise in pre-audit preparation.  I can support companies looking good when the state examiner comes.     

Issue - Part of the code was written in Cobol or another language not supported anymore.  And changes is needed.  Without SME around.  Can we feed this into a Large Language Model and get it written in a more modern language?

 

Solution - LLM are great for understanding code.  Not always efficient at writing new code.  But this could provide a good start for the re-writing exercise.  Don't forget the appropriate precaution to prevent public sharing of private information.

Issue - Confused on how to comply with upcoming regulations from Colorado, New York, and many other states? 

 

Solution - Following the twists-and-turns on the topic, I support companies developing their governance framework efficiently, inserting this new generation of models into the existing company internal model validation framework.

Issue - How to right-size the governance of AI models for your company, based on its size and risk appetite?  Creating a new C-suite position to head a new Board level committee is unlikely the right choice for a small to medium sized company.  

​

Solution - Leveraging my practical expertise, we can strike the right balance to allocate resources to this new requirements.

Issue - AI/ML models special characteristics trigger revisiting the model risk management framework.  And you don't understand where the wind is blowing, where NIST, NAIC Model Bulletin, Biden executive order, and various State regulations are pointing to?  How to make progress on the AI journey without risking major re-work when the regulatory framework crystalizes?    

​

Solution - I follow the twist-and-turns of the US regulation, Canadian and European regulation to understand where the wind is blowing.  Working assumptions can be easily articulated and reaffirmed once the regulatory thought process mature, allowing to reap the benefit of AI securely.