Mastering Actuarial Tasks with a Cloud Data & AI Platform

By Oliver Koernig, Shirly Wang and Suman Misra

Innovators & Entrepreneurs, September 2025

Introduction: The Evolving Role of Technology in Actuarial Work

Actuarial work in the life and health insurance industry is highly data-driven and complex. Actuaries perform a wide range of tasks such as experience studies, valuation, reserving, pricing, and risk modeling, all of which require vast amounts of data and sophisticated analytics. Traditionally, many of these tasks have been carried out using Excel and other legacy tools, but as data volumes grow and the need for more accurate predictions intensifies, actuaries are turning to more robust platforms.

Cloud Data & Artificial Intelligence (AI) platforms offer a unified platform that integrates data management, analytics, and machine learning, making it a powerful tool for actuaries. In this article, we’ll explore how Cloud Data & AI platforms can be used as an actuarial workbench, enabling actuaries in the life and health sector to streamline critical tasks like experience studies, valuation and reserving.

Key Actuarial Tasks in Life and Health Insurance

In this article, we will focus on three different actuarial tasks that can all benefit from using a Modern Cloud Data & AI platform:

Experience studies,
valuation and reserving, and
pricing and risk modeling.

We will also highlight some possible drawbacks of adopting such a platform and how to potentially address these.

1. Experience Studies: Data Integration and Analysis

Experience studies are crucial for life and health actuarial work, as they allow actuaries to evaluate past policyholder behaviors (such as mortality, lapse, and morbidity rates) to inform future pricing and reserving assumptions. These studies involve analyzing large datasets, often spanning years of policyholder information, claims data, and external factors like economic indicators. Due to the long-term nature of life and health insurance products, assumption setting becomes more critical to insurance modeling and financial projections.

Using a Cloud Data & AI platform, actuaries can:

Integrate and Govern Data: Experience studies require the integration of multiple data sources, including historical claims, policyholder demographics, and economic data. A central data catalog ensures that this data is centrally managed, governed, and accessible to authorized team members. It supports the integration of data from internal systems and external sources, streamlining the process of gathering all necessary inputs for the study.

Automate Data Updates: With the ability to schedule and automate data pipelines, experience studies can be updated in real-time or at regular intervals, ensuring that actuaries always have the most up-to-date information at their fingertips. This automation reduces the manual burden of importing and cleaning data, freeing up time for more value-added activities, especially as regulatory regimes like LDTI call for more frequent assumption updates.

Analyze Complex Patterns: With the rise of interest-sensitive and investment-linked insurance products, dynamic policyholder behaviors have become a reality. Actuaries need to analyze in-force movement in conjunction with market data, and potentially fit high-order functions to model the relationship. The ability to handle large volumes of data and support multiple programming languages provides actuaries with unprecedented flexibility compared to traditional tools.

Analyze Data with BI Tools: Once data is integrated, actuaries can use natural language driven dashboards to visualize trends in mortality, morbidity, and lapse rates. Dashboards offer interactive charts and graphs, allowing users to slice and dice the data, identify trends over time, and compare actual experience against pricing assumptions. For example, an actuary might track lapse rates over a 10-year period and visualize how they vary by age, gender, or policy duration, providing critical insights for adjusting pricing or reserving assumptions. Actuaries can create dashboards using natural language.

Natural Language-based query tools: These can also assist in simplifying data querying. Actuaries can ask questions like, “What is the average mortality rate for policyholders over 65 during the last five years?” and the tool will generate a visual or tabular response without the need for complex SQL or Python queries.

2. Valuation and Risk Management: Automating and Enhancing Accuracy

Valuation and risk management are fundamental actuarial functions, especially for life and health insurers. Actuaries must estimate the present value of future liabilities, ensuring that enough reserves are held to meet future claims. This involves complex calculations based on assumptions around mortality, morbidity, interest rates, and lapse rates. This is especially true for market-linked products like variable annuities, where stochastic modeling and dynamic hedging are involved. In addition, new regulatory regimes like LDTI and IFRS17 introduce several complexities to the valuation process such as more frequent assumption reviews and transition adjustment handling. Actuaries need tools that can not only perform big data computations but also manage operational complexities efficiently.

A Cloud Data & AI platform can help actuaries perform valuations through:

Improved Speed and Scalability: Valuation processes are often time-consuming due to the large volumes of data involved, particularly in life and long-term care products, where liability can extend over several decades. The ability to handle large datasets in a distributed computing environment ensures that even the most complex valuation tasks can be completed more quickly and with greater accuracy.

Valuation Process Management with Automated Workflows: Using automated workflows allows actuaries to define and manage complex valuation workflows with multiple dependent tasks, from loading in-force data, to post-process actuarial projection. This offers flexibility so that actuaries can set up tasks running at different frequencies (e.g., refreshing market assumptions monthly and policyholder behavior assumptions quarterly), as well as automation so that individual valuation tasks are executed in order (e.g., generating reports only after receiving up-to-date topside adjustments). The ability to view past run status also provides sufficient audit trails for disclosure requirements.

Reserve Estimation with Automated Machine Learning: While reserving calculations are often prescribed, machine learning can still help with reserve estimation in various ways. For example, when stochastic-over-stochastic calculations are involved, traditional actuarial projections are compute-intensive and time-consuming. Machine learning models can be used to derive proxy functions that estimate reserve requirements more efficiently. Using automated ML capabilities allows actuaries to set up machine learning models easily through a graphical UI and make machine learning an accessible tool among the actuarial community.

Scenario Selection: Advanced scenario selection techniques are another way to reduce stochastic valuation turnaround time. Automated ML allows actuaries to easily set up machine learning models to assess the severity of scenarios beyond traditional heuristics. As a result, instead of running stochastic-over-stochastic projections over thousands of scenarios, full projections are run on tail scenarios only. This significantly reduces the total run time required for valuation projections.

Enhanced Monitoring Capabilities: Dashboards can also be used to monitor key metrics related to reserves, allowing stakeholders to track the health of the reserve fund in real time. For example, actuaries can visualize reserve adequacy over time, broken down by product line or policyholder segment.

3. Pricing and Risk Modeling: Predicting Risk Exposure

Pricing life and health insurance products requires actuaries to model future risk exposure accurately. With the increased availability of data, actuaries are turning to more advanced techniques like predictive modeling and machine learning to improve the precision of risk predictions to inform better pricing and underwriting decisions.

Automated ML can be used to enhance risk modeling for pricing:

Predicting Claims Frequency and Severity: Using historical claims data and external factors (e.g., economic conditions, policyholder demographics), automated ML can be used to build predictive models that estimate the likelihood and severity of future claims. For life insurance, this involves predicting mortality rates for specific policyholder cohorts. For long-term care, it involves predicting the incidence and utilization of long-term care claims.

Risk Segmentation: Automated ML can also assist in identifying different risk segments within the policyholder population, helping actuaries set differentiated premiums that better reflect the risk profile of each segment. For example, an automated ML model might find that policyholders with certain lifestyle factors or health conditions are more likely to file claims, allowing the insurer to adjust pricing accordingly.

Modeling Catastrophic Risks: In addition to regular claims, life insurers must also account for catastrophic risks like pandemics or natural disasters. By integrating external data sources (such as economic forecasts or environmental risk factors), automated ML can help actuaries model the potential impact of these catastrophic events on their portfolios.

Possible Drawbacks when Adopting a Cloud Data & AI Platform

As with any solution, there are some possible downsides of adopting a cloud platform.

Here are a few examples.

The insurance company has not adopted the public cloud.
Data is not available in the cloud or is considered too sensitive.
Actuaries might not be familiar with cloud platforms and have mostly worked in a desktop/laptop setting using tools like Excel.

If a company has not adopted a public cloud like AWS, Microsoft Azure or Google's Cloud platform, then a Cloud Data & AI platform cannot be utilized. However, many companies who even a few years ago have stated that they won’t be going to the cloud have since changed their strategy, so the vast majority[1] of actuaries should have access already or will have access in the future.

Another challenge is that even if companies have adopted the cloud, the data may not be available, or it might be prohibited to use in the public cloud due to sensitive information.

In that case actuarial teams need to work with the security and compliance team to address these issues. Many cloud platforms have extended support for personally identifiable information (PII) and Health Insurance Portability and Accountability Act (HIPAA) information already and can support even the most stringent data security standards.

A third challenge is that many actuaries are accustomed to working in spreadsheets for tasks like pricing, reserving, and fine-tuning data. A mix of training offerings is a great way to allow the actuaries to learn new skills. Another way of bridging the gap between the power of Cloud Data & AI platforms and the familiarity of spreadsheets is a tool such as Sigma that offers a spreadsheet-like environment that integrates seamlessly.

A spreadsheet-like frontend provides a familiar interface for users while providing access to all the data entitled to them in the modern cloud data environment. Actuaries can view and manipulate any available datasets in the spreadsheet interface—or even create or import new data that can then be uploaded to the cloud. This even works with massive datasets that would typically overwhelm traditional spreadsheets.

A tight integration with the common data catalog obviates the conventional spreadsheet pitfalls of data replication and sprawl with a governed ecosystem that monitors data changes, tracks lineage, and is highly secure.

Conclusion: A Future-Proof Workbench for Actuaries

A Cloud Data & AI platform such as Databricks offers a robust and scalable solution for the actuarial tasks crucial to life and health insurers. From experience studies and reserving to pricing and risk modeling, a modern platform enables actuaries to work more efficiently and accurately. The integration of data management, BI tools, machine learning, and Excel-like environments ensures that both seasoned actuaries and insurance professionals new to advanced analytics can benefit.

By leveraging a Cloud Data & AI platform as an actuarial workbench, insurance companies can stay ahead of the curve in today’s increasingly data-driven industry, making better decisions, reducing risk, and enhancing profitability.

This article is provided for informational and educational purposes only. Neither the Society of Actuaries nor the respective authors’ employers make any endorsement, representation or guarantee with regard to any content, and disclaim any liability in connection with the use or misuse of any information provided herein. This article should not be construed as professional or financial advice. Statements of fact and opinions expressed herein are those of the individual authors and are not necessarily those of the Society of Actuaries or the respective authors’ employers.

Oliver Koernig, senior solutions architect at Databricks, located in New Jersey. His e-mail is oliver.koernig@databricks.com.

Shirly Wang, FSA & delivery solutions architect at Databricks, located in New Jersey. Her email is shirly.wang@databricks.com.

Suman Misra, senior solutions architect at Databricks, located in Oregon. His e-mail is suman.misra@databricks.com.

Endnotes

[1] https://www.capgemini.com/news/press-releases/91-of-banks-and-insurers-have-initiated-their-cloud-journey-yet-many-are-unable-to-realize-full-business-value/