GenAI and Annuities

November 30, 2024, marks two years since the release of OpenAI’s ChatGPT, which was the forerunner of other generative artificial intelligence (GenAI) models available to the public. Every day, about 200 million people use ChatGPT, and millions more use Claude, Bard, Gemini, Llama, or other large language model (LLM).

The use of AI is not new — it has been around in various forms for decades. But these LLMs represent a pivot point toward exponentially higher usage of AI, in the same way that the development of the World Wide Web and browsers opened the internet to the world. LLMs are designed to understand and produce human-like written output. They are trained on vast amounts of data, such as websites and books, to understand patterns and structures within language. By learning from these data sources, models gain the ability to respond to questions, generate coherent text, and even engage in real-time conversation through written and verbal modalities. The most advanced models have begun to demonstrate higher-level reasoning and mathematical functionality, with true artificial general intelligence on the horizon.

A central question for all industries is: How should they use GenAI? In what operations would GenAI produce an improvement? To what extent is GenAI applicable across all functions and lines of business? In terms of strategy, what are the business goals for GenAI integration, and how can they measure progress toward these goals? Given the pace of technological advancement, with older models quickly replaced by even more advanced versions, adjusting to and applying the new technology may feel like changing the tires on a moving vehicle. Yet companies cannot afford to be complacent, as other players in the industry could establish a major lead in GenAI adoption and experience all of the benefits that flow from it.

Annuities and GenAI

The annuity industry, like all businesses, has been trying to understand and make best use of this remarkable new technology for the past two years. In September, LIMRA asked 17 companies, representing nearly half of total annuity sales in the first two quarters of 2024, about their current and future GenAI usage within the marketing and distribution functions. Examples of marketing applications could include promotion and advertising, brand management, competitive intelligence, market research, and data mining. On the sales and distribution side, GenAI could be used for wholesaler support, client meeting preparation, illustration tools, product recommendations, order entry, or compensation administration.

As of 2024, the majority of annuity companies are not using GenAI for marketing-related functions within this business line. The most common usage was for competitive intelligence. These companies are occasionally, but not often, using GenAI to assist with the review of competitor’s materials, such as annuity product filings or announcements. Such quick scans and summaries of materials released by peer companies undoubtedly can improve efficiency by reducing the need for people to read lengthy technical documents.

Likewise, most annuity companies are not using GenAI as part of their sales- or distribution-related functions: Only two of the 17 companies use GenAI for wholesaler support or for client meeting preparation. No company reported using GenAI to help with product recommendations, order entry, compensation administration, or compliance reviews.

Measuring Success

GenAI is so new to the financial services industry that even establishing metrics for measuring progress can be challenging. Companies who have attempted to measure the effectiveness of GenAI on their annuity businesses’ marketing and distribution effectiveness tended to use efficiency-related metrics most often, followed by quality and cost metrics.

• Among companies who have measured effectiveness based on efficiency — how much faster the same task can be completed with the aid of GenAI — most claimed their company’s annuity marketing and distribution effectiveness improved only “slightly.”
• Companies who tracked quality enhancements — how much better tasks can be completed (i.e., a better end product, given the same inputs) — either saw only slight improvement or were not sure whether any such improvements were made.
• Measuring marketing and distribution effectiveness in terms of costs — how much cheaper tasks can be completed — was the least common metric used. Three of five companies using this metric reported slightly reduced costs, with two reporting GenAI had no impact or a negative impact on costs.

Despite the slow roll-out and limited success metrics, about one third of annuity companies surveyed plan to significantly or moderately increase their usage of GenAI for its distribution and marketing operations, with nearly half planning to increase its application slightly (Figure 1). No companies are planning to decrease GenAI usage. But four companies — about one quarter — were not sure how their usage might change over the next year, underscoring the many unknowns companies face as they set their plans.

Figure 1 — Expected Change in GenAI Usage Over Next 12 Months (Number of Companies)

All signs point to ever-increasing usage of this technology in our industry and society at large, but exactly how annuity companies integrate AI in the years to come will help to determine their long-term success. So far, the start has been slow, but the pace will only accelerate from here.

For additional details on this survey, see AI and Annuities: In Brief, LIMRA, 2024 (LIMRA log-in required).