Curious about how cloud and AI technology is going to impact the actuarial model migration process? We hope so, because we just covered this topic in depth in our most recent webinar - and we’ve got the summary ready for you.
In late August Montoux hosted Revolutionizing Actuarial Model Migration: Harnessing Cloud and AI for a Smoother Transition, a panel-style discussion between Montoux’s Klaas Stijnen, Meanwhile’s Jim Cristallo and CNO Financial’s Lindsay Quandt Obaji. Broadly, the webinar covered:
- The complexities of actuarial model migration and their origins
- How technologies like Montoux's cloud-native platform simplifies model migrations
- Best practices for initiating and streamlining actuarial model migration projects
- Post-migration support and success stories
Let’s dig into what we learned.
Complexities of actuarial model migration
First things first - why is it so important to consider actuarial model migration right now? Many life insurers must reconsider their business or blocks of business within a new economic reality and regulatory environment - and they need actuarial models to do it properly.
Unfortunately, many of these models need to be migrated to the cloud to be efficiently used and leveraged, and this is complex.
Here are some key reasons why:
- The number of models: some actuarial teams may be working with less than ten models, others may be juggling over 100, which dramatically impacts migration complexity
- Domestic or global: some companies are working with models and code which span geographies
- Documentation and code transparency: or should we say, the general lack thereof
- Your way or my way: there are as many ways to migrate a model as there are actuaries, and they all work slightly differently depending on preferences
- Lack of standardization: because model architecture and the migration process tends to be very inconsistent, the associated timelines and workflows become comparatively complicated
Technologies for simplifying actuarial model migration
When it comes to using AI technologies to simplify actuarial model migration, Montoux knows a thing or two.
“[Generative AI] is a great tool for productivity, but it can’t be relied upon blindly,” says Klaas Stijnen, Co-Founder and Co-CEO of Montoux. “It’s an incredibly capable assistant whose work needs checking. We always recommend, select your use cases based on those capabilities it has, but make sure you can check them.”
Why is Generative AI technology such a powerful tool for actuarial model migration? There’s a couple of key reasons.
It starts with the required compute power, which is massively increased because of the amount of data actuaries are working with today. This data must be cleansed, managed, and stored. It also needs to happen faster and more frequently than ever, while actuaries must simultaneously manage model and production outputs.
All of this creates a need for actuarial model migration to the cloud. Here are some ways Generative AI can help in this process:
- Translating one code language to another
- Building complex formula trees, which help actuaries understand the model and what they’re working with
- Documenting and explaining the new, migrated model very easily
Generative AI can significantly improve the migration process by serving as a model comprehension tool. Montoux has this technology built directly into its platform, automatically providing efficiencies to reduce time spent on model and code comprehension. This creates a more streamlined code and model development process while improving model risk management. It’s the difference between 4-5 pages of code versus 200.
“The biggest question is how to properly prompt Generative AI to do these things, which we’re doing a ton of R&D on at Montoux,” explains Klaas.
How to initiate and streamline actuarial model migration projects
The discussion moved on to how best to start a migration project.
CNO Financial’s Lindsay Quandt Obaji shares something she’s experienced within her own teams: “You have to find those hungry, creative, driven actuaries who can be the tip of the spear. These people are the ones who can create an appetite for change and really help teams start to phase out some legacy technologies.”
When it comes to a big migration project there are many different factors at play and teams must understand the technology you have to play with, including what it can and can’t do. Lindsay remarks that it's essential to have IT and actuarial teams on the same page to make this a reality and help overcome resistance to change.
Post-migration support and success stories
The conversation concluded with a discussion of what success looks like. The panel agreed journey transparency is vital, which is why it’s important to work with partners who don’t just hand over a model but ensure the end users understand and can leverage the migrated model.
“There’s a lot feeding into the models we migrate,” explains Jim Cristallo, an industry expert and Strategic Advisor for Montoux. “The clearer and broader your scope of understanding is about what you’re working with, and the earlier you can get it, the better.”
Klaas’ final recommendation for any actuarial model migration: “Create a definition of done - what does it look like when it’s fully migrated and what can we call a successful migration project.”
If you'd like to watch the full webinar recording, click here. If you’d like to make sure you don’t miss the next one, sign up for Montoux’s monthly newsletter. If you’d like to continue this conversation, we’d love to chat with you about it.