How generative AI can deliver value to insurers and customers
Insurers are struggling to grow their business and retain customers while managing profitability. They must comply with increasing regulatory burdens and compete with a wide range of financial services companies that offer investment products that can yield better returns than traditional life insurance and annuity products. Despite interest rates rising at an unprecedented rate over the past year as central banks attempt to curb inflation, a significant portion of insurers’ reserves are tied up in low-yield investments, which means investment returns are unlikely to improve for several years. portfolio is flipped).
Large, well-established insurance companies have a reputation for being very conservative in their decision-making and have been slow to adopt new technologies. They prefer to be “fast followers” rather than leaders, even when presented with a compelling business case. Fear of the unknown can negatively impact customer service and lead to project failures that result in losses.
IBM’s work with insurance clients, along with research from IBM’s Institute for Business Value (IBV), shows that insurer management decisions are driven by the need for digital alignment, core productivity and flexible infrastructure. To meet key challenges and transform their companies, insurers must provide digital services to customers, increase efficiency, use data more intelligently, address cybersecurity challenges, and provide resilient and reliable services.
To achieve these goals, most insurance companies have focused on digital transformation and modernizing their IT core through hybrid cloud and multi-cloud infrastructure and platforms. This approach can accelerate speed to market and improve the overall customer experience by providing enhanced capabilities for developing innovative products and services that help your business grow.
The role of generative AI in digital transformation and core modernization
Whether used for day-to-day IT infrastructure operations, customer-facing interactions, back-office risk analysis, underwriting and claims processing, traditional and generative AI are at the heart of core modernization and digital transformation initiatives.
Core Modernization with AI
Most major insurance companies have decided to migrate as much of their application portfolio to the cloud as possible as a mid- to long-term strategy.
When cloud use is combined with generative AI and traditional AI capabilities, these technologies can have a profound impact on businesses. The initial use of generative AI is often to increase productivity in DevOps. AIOps unifies multiple, separate, manual IT operations tools into a single, intelligent, automated IT operations platform. This allows IT operations and DevOps teams to respond more quickly (and even proactively) to slowdowns and outages, improving operational efficiency and productivity.
A hybrid multicloud approach combined with best-in-class security and compliance controls, such as those that IBM Cloud® supports for regulated industries, offers a compelling value proposition to large insurers across all geographies. Leading companies from every region partner with IBM on their core modernization journeys.
Digital innovation through AI
Insurance companies are reducing costs and delivering better customer experiences by using automation, digitizing their businesses, and encouraging customers to use self-service channels. With the advent of AI, businesses are now implementing cognitive process automation that helps enable customer and agent self-service options and automates many other functions such as IT help desk and employee HR functions.
The introduction of the ChatGPT feature has created a lot of interest in generative AI-based models. The base model is pre-trained on an unlabeled dataset and leverages self-supervised learning using neural networks. Foundational models are becoming an essential part of new AI-driven workflows, and IBM Watson® products have been using foundational models since 2020. IBM’s watsonx.ai™ foundational model library contains both IBM-built foundational models and several open source models. Hugging Face’s Large Language Model (LLM).
Supervised learning, which is used to train AI, requires a lot of human effort. It is difficult, requires intensive labeling and requires months of effort. Self-supervised learning, on the other hand, is computer-driven, requires little labeling, is fast, automated, and efficient. IBM’s experience with baseline models shows that labeling requirements are reduced by 10x to 100x and training times are reduced by 6x (compared to using traditional AI training methods).
To achieve digital transformation through AI, insurance companies must have a good understanding of their structured and unstructured data, organize it and manage it in a secure manner (while complying with industry regulations), and provide ready access to the “right” data. This capability is essential for delivering a great customer experience, attracting new customers, retaining existing customers, and gaining deep insights that can lead to new innovative products. It also helps improve underwriting decisions and reduces fraud and cost control. Leading insurers across all regions are implementing IBM’s data architecture and automation software in the cloud.
The generative AI capabilities that enable today’s digital transformation can be deployed in five areas:
- summary: Convert text from large documents, voice conversations, and recordings with domain-specific content into personalized outlines that capture key points (e.g., insurance contracts, policy and coverage documents, and responses to customer FAQs).
- classification: Reads and categorizes written input without any examples (e.g. triaging claims requests, triaging customer complaints, analyzing customer sentiment, triaging risks during insurance underwriting, analyzing customer segmentation for insurance product development).
- Generation: We create text content for specific purposes (e.g. marketing campaigns focused on specific insurance products, blog posts and articles on a variety of insurance-related topics, assistance drafting personalized customer emails, and generating code for use in insurtech systems).
- extraction: Analyze and extract essential information from unstructured text, such as extracting information from reports submitted by insurance agents or medical diagnoses from physician or clinical reports for use in insurance underwriting and risk assessment.
- question answer: Create question-answering capabilities based on specific data (for example, building policy- and coverage-specific Q&A resources for customer service agents).
There are many opportunities for insurance companies to create value as they begin to use generative AI for digital transformation of insurance business processes.
IBM’s collaboration with customers demonstrates significant productivity gains when using generative AI, including improving HR processes to streamline tasks such as talent acquisition and employee performance management. Increase productivity by freeing customer care agents to focus on higher-value interactions with customers (digital channel virtual assistants using generative AI handle simpler inquiries). Save time and effort modernizing legacy code with generative AI that supports code refactoring and transformation.
To take advantage of watsonx.ai, sign up for a free trial.
