(Image source: Visual China)

Lanjing News, February 18 (Reporter Shi Yu) - The ripple effects of DeepSeek are being felt across the financial sector, with insurance institutions swiftly accelerating their integration efforts.

Recently, several prominent insurance institutions, including New China Life Insurance, PICC Property and Casualty Insurance, and Taiping Life Insurance, have officially announced their partnership with the DeepSeek model. These integrations are primarily driven by the desire to enhance capabilities, boost efficiency, and optimize services. As a pioneering large model in China, DeepSeek is fueling the digital intelligence transformation and upgradation of the insurance industry. Industry insiders reveal that several other institutions are either "internally studying and discussing" or "advancing application plans across various lines," with some having already "quietly" integrated the model.

Amidst the digital intelligence transformation of the insurance industry, the DeepSeek large model is leveraging its open-source characteristics and low-cost advantages to provide differentiated enabling capabilities to industry institutions. Leading insurance companies are leveraging their prior technology investments and accumulated results to achieve further upgrades, while smaller and medium-sized institutions can utilize its low-cost training advantages to create customized solutions by integrating local data with large models, thereby breaking through technical cost barriers. The application of DeepSeek in the insurance industry is poised to elevate the overall industry's digital intelligence level, potentially redefining benchmarks.

Insurance Institutions Embrace DeepSeek, Encompassing Business Expansion, Claims Processing, Investment Research, and Beyond

As the wave of DeepSeek spreads, the insurance industry is stepping into this new era.

New China Life Insurance took the lead in officially announcing the successful integration of two model products, DeepSeek R1 and V3, into the New China eHome App. By leveraging the App's underlying technology stack and the interoperability of the Internet technology ecosystem, the company has created personal AI assistants that support multiple intelligent application scenarios.

Subsequently, insurance companies such as Beijing Founder Life Insurance, Taiping Life Insurance, and PICC Property and Casualty Insurance, insurance asset management firms like PICC Asset Management and Taiping Asset Management, and insurance intermediaries like Tai Chi Hua Bao's Baolianke have also officially announced their integration with DeepSeek, actively embracing this trend. Additionally, this reporter learned from industry sources that some institutions have already conducted internal tests and integrations, while others are "internally studying and discussing" or "advancing application plans across various lines."

From a practical standpoint, focusing on business expansion scenarios, Beijing Founder Life Insurance officially launched the intelligent business expansion assistant "Fangling" based on the DeepSeek large model on February 14. Initially, the "Fangling" assistant will unlock the basic function of "intelligent knowledge retrieval," serving as an E-marketing intelligent assistant that provides comprehensive answers covering basic insurance concepts and other content for agents. Following this, knowledge training will commence, with a simultaneous emphasis on systematic training for agents to utilize AI intelligent tools.

In the claims processing segment, PICC Property and Casualty Insurance has fine-tuned the DeepSeek model to support automated claims quality inspection, accurately extracting key information such as case timeliness and damage locations from claims materials. This enables intelligent coaching for claims personnel, actual testing of intelligent Q&A, and the provision of improvement suggestions based on current personnel skills.

In the realm of investment research, PICC Asset Management indicated that it recently completed the local deployment of the DeepSeek R1 large model and created a dedicated interactive application accessible to all company employees. Concurrently, it is exploring deep integration of DeepSeek with application systems to realize intelligent research report scenarios.

Taiping Asset Management has also recently completed the local deployment of the DeepSeek-R1 large model and achieved localization adaptation. By deploying DeepSeek locally, it can offer content services such as industry trend analysis and local data interpretation in areas like element extraction, intelligent retrieval, and knowledge Q&A. The company further proposed that in the next phase, it will delve deeper into the model's capabilities, gradually attempt applications in the fields of investment, operations, and risk control based on actual work scenarios, and leverage advanced "AI" technology to propel the development of asset management towards intelligence and informatization.

Harnessing DeepSeek's Open-Source and Cost-Effective Nature to Accelerate Industry Digital Intelligence Upgrades

Beyond the current layout, various insurance institutions have outlined further plans.

For instance, Beijing Founder Life Insurance envisions that in the future, intelligent assistants will evolve into more scenario-based application tools, gradually introducing "AI intelligent coaching" and "underwriting assistants," among others, to construct a comprehensive support system encompassing pre-, mid-, and post-business stages. Taiping Asset Management, on the other hand, plans to further explore the model's capabilities, gradually attempting applications in the fields of investment, operations, and risk control based on actual work scenarios, and leveraging advanced "AI" technology to drive the intelligence and informatization of asset management.

Regarding the application of DeepSeek, the industry anticipates upgrades across the entire chain. DeepSeek has provided "official" answers regarding the directions and scenarios in which it can empower the insurance industry, encompassing intelligent customer service and education optimization, underwriting and claims automation, data-driven product design and actuarial science, customer lifecycle management, compliance and operational efficiency enhancement, investment and asset management, technological advantages, and landing scenarios, among other dimensions.

Deployment is not an overnight achievement, nor are plans mere fantasies. In fact, insurance institutions have collectively been driving digital intelligence transformation in recent years, with the implementation of DeepSeek further accelerating the industry's transformation and upgrading based on their existing foundations.

Leading institutions have actively invested in the AI field and large model construction, and DeepSeek's open-source characteristics and high performance provide robust underlying support and upgrade enablement for their accumulated results. For example, PICC Property and Casualty Insurance officially announced on February 18 the introduction of the full-size DeepSeek large model, completing privatized deployment based on domestic computing power equipment, integrating DeepSeek capabilities into the company's large model foundation, and constructing the professional service capabilities of the DeepSeek large model for insurance business based on the insurance business data set accumulated during previous explorations of AI technology.

For small and medium-sized insurance institutions, they benefit from DeepSeek's low-cost training and open-source characteristics, accelerating their catch-up in digital intelligence levels. By combining local data with large models, they can create customized models, eliminating certain cost barriers. Overall, the industry's digital intelligence level is expected to be enhanced through the application of DeepSeek.

"Compared to traditional AI, DeepSeek boasts stronger data processing capabilities and a higher level of intelligence. It can more accurately analyze and predict customer needs, offering personalized insurance products and services. Moreover, DeepSeek's self-learning ability makes it more efficient in handling complex insurance cases, enabling it to swiftly adapt to market changes and the diversity of customer needs," Lan Jie, General Manager of the Wisdom Delivery Center at Yuanxin Huibao, analyzed to this reporter. These advantages give DeepSeek a significant edge in enhancing insurance companies' operational efficiency and customer service quality.

It is noteworthy that Lan Jie emphasized, "During the deployment of DeepSeek, insurance institutions must prioritize data security and privacy protection. This includes ensuring the encrypted storage and transmission of data and adhering to relevant data protection regulations. Additionally, the application of the model requires rigorous testing and validation to ensure its accuracy and reliability."

Dialogue Between Frontline Marketing Personnel and DeepSeek: Empowerment or Replacement?

Unlike the concerted efforts of insurance institutions to integrate DeepSeek, frontline insurance marketing personnel have a more direct and adaptable approach to "integrating" DeepSeek. In fact, with the rise of AI and the advent of ChatGPT, the debate over whether artificial intelligence will replace insurance agents has lingered for some time, and the emergence of DeepSeek has added another layer of uncertainty to this discussion.

"I have used DeepSeek to provide protection recommendations and product suggestions based on clients' individual circumstances. The output content is more detailed than previously generated by AI, and the specific display of its thinking process helps the inquirer understand more clearly why these protections are necessary and the differences between various product types," Fang, an insurance broker from a leading insurance intermediary agency, told this reporter. "However, this is merely the initial step in the sales process. Clients often struggle to know which aspects to further compare and might even blindly choose products based on AI's relatively 'universal' recommendations."

What cannot be overlooked is the "hallucination" phenomenon associated with AI models, including DeepSeek, which necessitates users to conduct multi-source verification and screening. However, given the high professional barriers and complex clauses of insurance products, this undoubtedly increases the difficulty for users.

"Insurance sales are inherently human-centric, often built on trust and understanding," a business leader from a branch of an insurance intermediary agency told this reporter. "Furthermore, customer needs are diverse, and insurance clauses are intricate. Professional insurance salespersonnel need to match and make decisions based on their own experience and comprehensive understanding of customers and products, which cannot be achieved through simple searches."

"But the help is undoubtedly greater than the threat," the aforementioned business leader from the insurance intermediary agency branch summarized. Regardless of whether it poses a threat, embracing DeepSeek has become a consensus among frontline business personnel.

Judging from the feedback of frontline agents, DeepSeek can assist them in generating basic suggestions and solutions for clients, enabling them to optimize these for enhanced efficiency. It can also offer marketing tips, engage in content creation, etc., freeing up basic service energy and further shifting the focus of work and business expansion models towards providing professional risk protection solutions.

Just as the direction of AI cannot be fully predicted, how the insurance industry will harness AI and DeepSeek to achieve transformative changes remains to be seen over time.