Montoux Generative AI How To Start Instructions
- July 24, 2024
- Montoux
Table of Contents
- Montoux Generative AI How To Start Instructions
- How to Start with Generative AI
- Identify potential high value actuarial AI use cases by adopting a
- a) Foundational Awareness
- b) Identify Potential Use Cases
- c) Assess Feasibility and Impact
- d) Prioritization Matrix
- Identify AI champions – who leads the adoption of AI at organizational,
- Get ahead of security, quality and scalability
- Partner for success
- Read More About This Manual & Download PDF:
- Read User Manual Online (PDF format)
- Download This Manual (PDF format)
Montoux Generative AI How To Start Instructions
How to Start with Generative AI
Montoux is working with leading life insurers around the world to identify,
prioritize and implement generative AI use cases with the primary goal of
making actuaries more productive.
Recognized as a thought leader on the topic of how generative AI will impact
actuarial work, Montoux has run a webcast for the Society of Actuaries as well
as released blogs and articles on the topic.
The following document is a simple 4 step guide to help actuarial teams get started on their gen AI journey:
Identify potential high value actuarial AI use cases by adopting a
framework like the following
(the example framework was created by AI)
a) Foundational Awareness
Create foundational awareness of Generative AI, including an explanation of how Gen AI can be relevant to actuarial work. The following are examples of good, free foundational courses in Gen AI:
- Generative AI for Business Leaders
- Generative AI Learning Plan for Decision Makers
- The AI Bootcamp
- AI Foundations for Everyone Specialization
- Generative AI for Everyone
- Generative AI Fundamentals
- Introduction to Generative AI
b) Identify Potential Use Cases
Run idea generation workshops and categorize potential use cases into ie. model development and governance, data augmentation (creation of synthetic data sets for modeling), predictive modeling, and reporting (automating report generation and data analysis).
c) Assess Feasibility and Impact
Technical Feasibility: Evaluate the technical requirements for each use case,
including data availability, model training, and integration with existing
systems.
Business Impact: Assess the potential impact of each use case on key business
outcomes, such as improved risk assessment, cost reduction, enhanced customer
experience, and compliance.
d) Prioritization Matrix
Value vs. Feasibility Matrix: Create a matrix to rank use cases based on their
business value (impact) and feasibility (technical and operational). This
helps in identifying high-impact, high- feasibility projects for immediate
focus.
Criteria for Ranking: Include criteria such as potential ROI, time to
implementation, alignment with strategic goals, and resource requirements
Identify AI champions – who leads the adoption of AI at organizational,
department and team levels?
Ernst and Young suggest that organizations should “put humans at the center of Gen AI strategy to succeed”.
Actuarial departments should identify the key people who will lead adoption of AI. We have developed this simple survey that can be distributed around actuarial teams to identify team members who have the greatest interest in and potential for being AI adoption leaders.
- On a scale of 1 – 10, how transformative do you think AI will be to the actuarial profession?
- On a scale of 1 – 10, how much do you think AI will affect your career?
- On a scale of 1 – 10, how would you rate your belief in AI as a catalyst for change within your organization?
Get ahead of security, quality and scalability
Gen AI PoCs and experiments are a good way for an organization to understand capabilities, however our experience is that shifting from PoC phase to production can be a challenge. To avoid the risk of a successful PoC not translating into a successful production solution (and to reduce the time it takes to achieve this), we recommend that actuarial departments consider the following questions as they assess a potential solution:
- Does the proposed AI solution meet the organization’s enterprise security and data privacy requirements? g. SOC2 + make specific data governance references
- Is the proposed AI solution supported post-PoC, i.e. can it be successfully deployed in production? How is this evidenced?
- How does the proposed solution manage quality and model accuracy?
- Is the proposed solution likely to be scalable and able to be integrated with the existing technology ecosystem? Is the technology open or closed in nature?
- How ready is the organization to procure Gen AI technology – are there standard contract terms or templates, etc
Partner for success
Some degree of technology partnership is likely to be a requirement for actuarial departments seeking to implement Gen AI. This may be a partnership with a purpose-built Gen AI application provider like Montoux, a cloud infrastructure provider like AWS, or an internal partnership with colleagues in technology teams. Making the right partnership choice is dependent on the dynamics of the specific actuarial team. It’s worth considering the following questions with regard to partnership strategy:
If pursuing self-build Gen AI solutions, what skill set does your team or organization possess? Montoux’s experience is that successful Gen AI solutions require a mixture of technical actuaries, AI engineers for retrieval augmented generation (RAG) and data structuring, prompt engineers, and application engineers (to productize the solution).
- If considering a partner solution, can the partner evidence credibility through demos and case studies?
- Has the partner/vendor produced relevant industry thought leadership content?
- Does the partner align with your actuarial team’s culture and ways of working?
With Gen AI, seeing is believing. We strongly encourage actuarial leaders to create an environment whereby Gen AI curiosity is fostered and teams are encouraged to experiment (guided by a framework like the one outlined in this guide). Our belief is that the 4 simple steps provided in this guide will prepare actuarial departments well to commence a successful Gen AI journey.