By now, you are probably already familiar with the perspective of large language models (LLMs). These augmented intelligence tools with user-friendly interfaces combine artificial intelligence (AI) with human intelligence to enhance and expand human capabilities: content creation, quick responses, decision making, problem solving, and more. LLMs will drive digital transformation for organizations that use it, prompting reflection on the value and talent requirements of human staff.
The news of this revolutionary technology, which was largely made known by the launch of ChatGPT by OpenAI and Microsoft in November 2022 and the subsequent launch of the Google Bard AI chatbot, has captured the attention of the world – and for good reason. LLMs represent the next generation of Natural Language Processing (NLP) and Natural Language Generation (NLG). Using billions of parameters and diverse public datasets, deep learning and transformation based LLM models provide flexible, diverse, public outputs with vendor-defined limits.
What impact can LLMs have on life insurance in particular?
Value
Current LLMs offer two main categories of value in life insurance, helping different end users and use cases:
- “Tell Me” (descriptive) provides functionality for both customers and employees. At this fundamental level, LLMs can provide training and guidance. For example, they can offer policy information (providing policyholders with quick and accurate information about their insurance coverage, deductibles, and other policy details) or text synthesis and analysis (pulling from various documents and learning from organizational information to identify specific elements). This can be useful with unstructured information commonly found in the realm of life.
- “Do It For Me” may refer to smart bots for customers that provide information and troubleshooting, or employee services such as risk mitigation tools. LLMs can help manage everything from marketing (content generation, social and email marketing, data analytics and A/B testing) to insurance-specific processes related to underwriting (gathering information about a candidate to determine a risk profile; analyzing an integrated health care, insurance and alternative services). data for direct expedited underwriting and claims processing with automated initial steps to collect policyholder information, complete data entry and document verification, and determine eligibility.)
LLMs offer the potential to provide even more life insurance value in two additional categories. Together they can analyze large amounts of data (for example, to identify fraudulent data; combine health data with policy data to predict potential fraud; detect patterns and anomalies), improve customer service (by integrating into consultant applications and websites to provide instant responding to customer inquiries, reducing the burden on consultants) and improving operational efficiency for various insurance functions (including claims processing, fraud detection, underwriting and premium calculation).
- “Tell me” (predictive), unlike the descriptive functionality above, this category relies on the potential of finely tuned generative pretrained transformer (GPT) models. Pre-training on large amounts of text data allows the model to learn patterns and relationships in the data, fine-tuning that information for specific language tasks (such as text generation, question answering, and sentiment analysis).
- “Advise me” covers the field of machine learning (ML) and a sub-field of deep learning models such as decision transformers.
economy
LLMs offer the potential to enable a new wave of automation, risk/loss mitigation and data-driven decision making for insurers. Payback was made possible by improved market data analysis to better inform growth strategy and the ability to leverage predictive customer behavior modeling. The cost of these initiatives is related to maximizing automation and minimizing risk/waste.
Income can be increased through direct (product sales) and indirect (share of wallet income) opportunities. LLMs can earn direct income through pricing optimization; orient current customers to optimal products; improving the product selection experience for new customers; and enhancing the strategic and creative work of insurance staff, removing repetitive tasks and providing greater customer insight. Indirect income can be generated by improving the customer experience, providing personalized recommendations and creating upsell and cross-sell opportunities for existing customers, as well as increasing engagement and brand awareness for new customers.
Celent foresees rapid adoption of LLM by businesses in the coming years. In 2023-2024, early adopters (~20%) will be innovation-driven companies experimenting, tweaking, and building/testing LLM prototypes. 2025-2026 will be a critical review period (relative to economics, compliance) during which early majority (21-50%) will invest in LLM and late majority (51-75%) will adopt and integrate LLM around 2027-2026 . 2028. LLM is likely to reach maturity by 2029, when the laggards (76%+) will adopt the technology and when use case growth is likely to slow down.
Implementation of the strategy
Where on the curve will you be? Insurers seeking to adopt and implement an LLM strategy must be forward-thinking, courageous and responsible. They must evaluate their short, medium and long term goals; the need for dedicated resources; and cost implications for platforms and services. They must be willing to accept disruption and paradigm shifts. Their development and deployment of AI must be accountable, appropriate, ethical, impartial and transparent. This requires steps across the organization as the implementation of an LLM strategy can affect the technical infrastructure, business model, operating model, and culture of the insurer.
Because LLMs are relatively new and rapidly evolving (including, at the time of writing, the launch of the latest OpenAI model, GPT-4), insurers will need to consider their policies and methods for deploying these technologies. For example, filters or other measures to limit the use of the LLM for certain initiatives may be appropriate. Careful consideration of the LLM is also important, for example, in order to navigate the ethical considerations of bias.
The pace of LLM innovation can be dizzying, but inaction comes with its own risks. Competitors using LLMs can gain an advantage that life insurers who decide to wait will find it difficult to close. Clients who have used an LLM, perhaps in a different industry, will have higher service expectations, such as AI-powered chatbots. Employees who value AI/ML tools—especially because they allow employees to focus on value-adding work—may leave for other, more advanced firms. Last but not least, life insurers that continue to rely on manual processes may suffer from inefficiencies that LLMs could automate, reducing operational efficiency and increasing costs. While each represents a potential risk, each also provides an opportunity for innovation and success for those life insurance companies that choose to invest in unproven technology and dictate the pace of change.
