The recent unveiling of the Chat-Trust 3.0 by Ximei Mutual Life Insurance marks a significant milestone in the application of artificial intelligence (AI) within the insurance sectorThis product is touted as a vertical application of large language models specifically tailored for the intricacies of insurance, signaling a shift towards more intelligent, responsive customer service and claims processing in this traditional industry.

According to a recent report from the International Data Corporation (IDC), the usage of enterprise-level AI is projected to surge by 25% in 2024 compared to 2023, with spending in this field skyrocketing by a staggering 250%. This indicates a growing recognition of AI’s potential to enhance efficiencies and improve service offerings within various industries, particularly insuranceAs AI matures, its applications have progressed from mere support functions within companies to pivotal roles in underwriting, claims processing, and product analysis, demonstrating a profound impact on the insurance industry's evolution.

Different insurers are adopting unique strategies when it comes to integrating AI into their operations

Advertisements

Tellingly, Ximei Mutual Life’s digital operations chief, Cheng Yu, has highlighted that rather than incurring high costs associated with building specialized models, their approach is founded on low-cost, rapid iterations to address specific industry challengesBy leveraging general foundational models, they emphasize enhancing the management and processing of unstructured dataThis includes the use of retrieval augmented techniques and model fine-tuning, allowing them to carve out specific applications tailor-fit to the insurance landscape.

In contrast, ZhongAn Insurance, which has built its reputation on technological investments, straddles dual priorities of developing vertical applications and exporting technologyIn 2023, ZhongAn launched a generative AI middle platform named "Lingxi," along with insurance-specific applications including an intelligent content creation platform called "Yichuang" and an upgraded AI version of its operational analysis platform, "Jizhi." The company reported impressive growth, with total revenue from technology exports reaching 829 million yuan, reflecting a 40% increase year-on-year.

The influence of AI extends heavily into the realm of insurance sales as well

Advertisements

Huize, a significant player in the sector, reported that they significantly enhanced operational efficiency through digital means, continually refining algorithms to boost advisor productivityTheir "AI Marketing Assistant" is deployed broadly across their sales consultants, while the launch of its "Business Intelligent Agent" is set to assist in the digital transformation of the industryThe CEO of Huize, Ma Cunjuan, stated that the company intends to accelerate the deployment of its proprietary AI models across the insurance service chain, fully empowering insurers, agents, and sales channels alike.

Unlike typical large models that operate on a “black box” basis—where inputs yield outputs without transparency—Ximei Mutual Life has pioneered a “white box” model in its latest iterationIn the application of AI within the insurance context, black box models can efficiently provide answers but do so with a level of obscurity that leaves users in the dark regarding the underlying rationale

Advertisements

On the other hand, the white box approach employed by Ximei allows for an illuminating glimpse into the reasoning processes behind the outputsFor instance, when assessing a health insurance product, their model meticulously details how data such as policyholder age, medical history, and family genetic predispositions interconnect within intricate algorithms and expertise in insurance to arrive at premium rates and coverage decisions.

According to Tong Guohong, head of the data information center at Ximei, insurance products inherently involve complicated risk assessments and decision-making proceduresThe implementation of a white box model not only fosters greater clarity but also aids in regulatory compliance, as it allows supervisory bodies to easily grasp the logic underpinning assessmentsThis enhanced understanding can facilitate more effective compliance reviewsMoreover, clients benefit from transparent reasoning processes that cultivate trust in policy terms and pricing

• Digital Transformation Boosts Consumer Goods Efficiency
• Banks Turn to Volume Amid Interest Rate Cuts
• Accelerating AI Adoption in the Insurance Industry
• U.S. Inflation Falls Short of Targets
• Market-Driven Debt-to-Equity Swaps Spur Innovation

Internally, the insights gleaned from this transparency help teams continuously optimize models and operational workflows.

Industry experts emphasize that as the application of large models proliferates, stringent controls must be established to ensure data securityWith user information safety being paramount, it’s crucial that every interaction node within the architecture interfacing with the model strictly implements user privacy measuresThis involves severing the connection between user confidential information and the model itself, creating robust defenses for personal data protectionFor example, encrypting data during transmission is essential to prevent interception or tampering; similarly, employing block storage for user privacy data with multiple layers of access restrictions ensures that only authorized personnel or processes can reach critical information, thus enhancing the safety of AI applications in the insurance landscape.

This layered approach to privacy safeguards reflects the broader industry trend towards maximizing AI's advantages while safeguarding user trust