On October 27th, the "Second (2024) Insurance Technology and Intelligent Data Conference" was held in Beijing, hosted by the China Insurance Association, guided by the People's Government of Tongzhou District, Beijing, and jointly organized by the China Insurance and Retirement Financial Research Center at Tsinghua University PBC School of Finance, Beijing Institute of Financial Technology, and Sunshine Insurance Group. The conference was attended by over 350 executives, experts, and scholars from 173 insurance companies, insurance technology companies, higher education institutions, think tanks, etc., discussing the theme of "Improving the level of insurance digitization, enhancing industry's sustainable development capability," focusing on the practical applications of big data and artificial intelligence in the insurance field, grasping new technology trends, innovating business models, and creating a brand new future.

Waterdrop Inc. founder and CEO Shen Peng was invited to attend the conference and delivered a keynote speech. Shen Peng stated, "The breakthrough in the 'big model' technological capability brings new opportunities for the high-quality development transformation of the insurance industry, and insurance technology is entering the era of 'intelligence and digitization.'

Image: Waterdrop Inc. founder and CEO Shen Peng delivering a speech

Insurance Industry Enters the Era of Intelligence and Digitization

Shen Peng introduced that the digitization process of the insurance industry has gone through informatization, onlineization, and has now entered the era of intelligence and digitization. The deep integration of big model technology with the insurance industry will have a profound impact on the industry's industrial structure. The advantages of the insurance industry will gradually shift from channel advantages, product advantages, agent scale advantages, etc., to advantages empowered by technology applications.

In the past decade, the scale and efficiency improvement of the insurance industry have not been significant, and the essence behind this is that the proportion of technology in the insurance service chain is not high. Taking the sales service end as an example, this is a high-cost link in the insurance industry, but traditional insurance sales lack much technological support. The large model's "contextual semantic understanding ability," "natural coherent language expression ability," and "learning ability to deal with new scenarios" can greatly assist salespeople, which brings great potential for cost reduction and efficiency improvement.

Taking Waterdrop as an example, under the empowerment of the large model throughout the process, Waterdrop Insurance's customer acquisition capability has been systematically upgraded, and human efficiency has also been scaled up. A team of about 10 people can even leverage hundreds of millions of annual premiums.

Self-developed AI insurance brokers' assistant surpasses junior service personnel in service efficiency.

Shen Peng introduced that since 2019, Waterdrop has started to invest in AI construction, with an annual R&D investment of around 0.3 billion yuan, and has built a research and development team of several hundred people, with core team members from top universities such as Peking University, Tsinghua University, and the Chinese Academy of Sciences, and has obtained dozens of technical patents in the field of AI. Based on long-term technological investment, Waterdrop has independently developed a large model in the insurance vertical field - the Waterdrop Shuishou Grand Model, which has been applied in all scenarios such as sales assistance, intelligent marketing, intelligent quality inspection, and product innovation.

Image: Waterdrop's self-developed insurance field large model - Waterdrop Shuishou Grand Model

At the sales assistance level, based on the empowerment of the "Waterdrop Shuishou Grand Model," the AI insurance brokers' assistant "Bang Bang" has undergone a deep upgrade. Previously, "Bang Bang" mainly undertook repetitive and labor-intensive work in the insurance service process, and after the upgrade, "Bang Bang" can take on more complex tasks in the sales planning process, providing more professional, accurate, and understandable advice for insurance brokers and users, with service efficiency surpassing junior service personnel.

Without the empowerment of AI large models, the cost of timely answering various complex user questions is very high. After the large model appears, the marginal cost of insurance service is very low, and the large model can accurately and quickly answer users' questions, greatly enhancing user satisfaction. For example, many users may have questions when purchasing insurance. Have they undergone brain surgery, or have gestational diabetes or other illnesses, can they buy a certain product. It is difficult to find the answer just by looking at the insurance application page, and users may interrupt the insurance purchase process. In this scenario, "Bang Bang" will answer various underwriting, claims, premium, and other questions for users.

From a marketing perspective, waterdrop launched a tool called "Waterdrop Microvid" this year, which empowers insurance brokers by creating digital avatars through AI training, reducing the traditional several hours of video shooting time to just a few minutes. A collaboration between Waterdrop Microvid and a well-known insurance company's insurance brokers showed that within a month, through just one channel, the video achieved over 0.5 million views, over 0.03 million interactions, and an increase of over 2800 fans, greatly enhancing the marketing efficiency of the new media channel.

The AI's advanced contextual semantic understanding and long-text inference capabilities play a vital role in quality inspection. Currently, waterdrop's intelligent quality inspection can achieve 100% comprehensive coverage, including operations across channels such as voice, enterprise WeChat chat records, and Moments. After continuous training, the accuracy of intelligent quality inspection has nearly reached the average level of manual inspection. The cost of quality inspection has also significantly decreased, with AI quality inspection costs over 50% lower compared to manual quality inspection costs when factoring in the cost of manual review.

Furthermore, in terms of product innovation, leveraging AI technology and big data advantages, Waterdrop Ins can better understand the insurance needs of segmented users such as those with illnesses, the elderly, and pregnant women, collaborating with insurance companies to drive supply-side innovation and customize more cost-effective insurance products.

Shen Peng stated: "In the future, Waterdrop will continue to enhance its specialized operational capabilities, build a professional service team, and provide better services to customers. At the same time, it will also continue to strengthen the exploration of AI large models and other data intelligent services, gradually opening up to the industry and working with partners to promote the high-quality development of the insurance industry."