RNA Analytics Accelerates Insurance Actuarial Industry's “Digital Twin” Transformation with Full-Scale AI Project

Authored by Justin Hwang, Head of the AI Project and Graham Howarth, Chief Technology Officer

RNA Analytics, a leading provider of actuarial and risk management solutions, has officially launched a large-scale project to revolutionize actuarial modeling through AI, reinforcing its position at the forefront of digital transformation in the insurance industry. This initiative aims to dramatically improve the labor-intensive aspects of actuarial modeling by integrating AI-powered solutions with the company’s core R3S software platform.

Through this project, RNA Analytics plans to automate the repetitive modeling and data preparation tasks traditionally handled by actuaries, allowing them to focus more on analysis and strategic decision-making. This long-term vision, referred to as the “Digital Twin Transformation,” begins with a “Quick Win” phase targeting product development functions and will gradually expand to all areas of actuarial work.

The AI project centers around three core functionalities: First, when there are changes to base documents (e.g., assumptions), which in turn affect actuarial outputs such as premiums and reserves, users can upload the revised base documents to the R3S AI engine. The system will then automatically create or update the corresponding “Product Information Tables.”

At the same time, the AI engine works within the R3S Modeler—RNA’s cash flow modeling tool—to automatically generate models for premium and reserve calculations as well as future cash flows. This ultimately leads to the creation of profitability analyses and Present & Value (P&V) tables.

[Figure 1] Example Workflow of Using the R3S AI Solution for Product Revision Tasks

The second core component is the R3S modeling support system, powered by RNA’s proprietary actuarial-focused language AI model, Acty LLM. Initially launched in the form of a chatbot, it allows R3S users to search for help, ask questions about actuarial theory and practice, and inquire about relevant regulations. Over time, it is expected to evolve into a full-fledged Modeling AI Agent, capable of generating, auto-filling, and validating code. Ultimately, RNA aims to enable fully automated actuarial modeling functionalities that meet regulatory requirements such as IFRS 17 and Solvency II (K-ICS).

The third pillar involves defining an ontology[1] of core tasks in product development teams to maximize the applicability of AI. Based on this ontology, RNA plans to transition the entire insurance product development process into a digital twin, enabling strategic decision-making through extensive simulations.

The workflow for AI-driven agents is also being developed in phases. Knowledge-based Q&A agents, user support agents, and code generation/validation agents are scheduled for completion within the year. Additionally, RNA plans to introduce automated report generation agents in future phases.

[Figure 2] Agentic Workflow Architecture of the R3S AI Solution

For this project, RNA Analytics has built dedicated datasets and developed its own proprietary LLM specialized in actuarial science. The company is also accelerating efforts to offer both cloud-based and on-premises service options.

Commenting on the AI products being developed through this initiative, Justin Hwang, Deputy CEO and Head of the AI Project, stated: “While the initial release will focus on automating core actuarial tasks, like many other AI products, the learning curve of the LLM will rapidly accelerate as user feedback accumulates—eventually evolving into an extremely intelligent AI agent.”

Meanwhile, Graham Howard, Managing Director and Chief Technology Officer, emphasized: “This AI project is designed to improve the productivity and accuracy of actuarial modeling processes, providing real, practical support to actuarial professionals. Ultimately, our long-term strategy is to enable a full actuarial digital twin that can simulate complex workflows and leverage the vast data resources held by insurers.”

[1]Ontology: a formally defined system of concepts, properties, relationships, and rules within a specific domain.