A new trend is emerging as the global financial industry rapidly adopts artificial intelligence. AI is no longer just a backend tool but is now taking the form of "Intelligent Agents" that directly participate in customer service, decision support, and even business execution. Banks and securities firms have been experimenting, and now insurance companies are joining in.
Traditionally, purchasing a life insurance policy often meant facing delays for simple tasks like checking policy details or inquiring about surrender values, as these required manual responses. Questions such as "How much premium have I paid?" or "What would I get back if I surrender now?" seem straightforward but involve personal data and complex calculations, making客服 cautious and customers impatient. This gap between long-term commitments and immediate service has long been a challenge for the life insurance sector.
China's largest life insurer, China Life Insurance, is addressing these frequent yet sensitive queries through its overseas platform using a new type of AI known as AI Agent. The company has proactively partnered with NetEase Zhihui's cloud business to deploy AI Agent applications in after-sales services. Official disclosures indicate that the system achieves over 90% accuracy in handling scenarios like policy maturity and surrender consultations. For instance, instead of manual checks, the AI Agent now automatically retrieves data, computes results, and provides compliant responses, significantly reducing wait times.
The core issue for life insurers lies in the complexity behind these seemingly simple customer inquiries. Post-sale services—such as policy checks, payment queries, and changes—require precise calculations based on individual policy details, including payment history and附加 benefits. Manual processing is not only time-consuming but prone to errors, which can lead to customer complaints or disputes. For example, calculating the cash value of a policy surrendered after years of payments involves accessing multiple systems and applying actuarial rules, where human oversight can easily result in inaccuracies.
An AI Agent, or artificial intelligence entity, differs from conventional chatbots by actively understanding goals, planning steps, utilizing tools, and completing tasks autonomously. When a customer asks about surrender options, the Agent identifies the intent, fetches policy data, computes values, and explains procedures without human intervention. However, in financial contexts, this is challenging due to strict compliance and data privacy requirements. AI cannot operate freely; every action must adhere to predefined rules, such as avoiding disclosure of unauthorized information or making decisions on behalf of clients. Thus, financial AI Agents function like "professionally trained digital employees," knowing when to act, escalate, or defer to human judgment. In practice, systems still transfer complex cases to human客服 to balance safety and experience.
This is not China Life (Overseas)'s first foray into AI transformation. Earlier this year, the insurer launched two AI applications in Hong Kong and Macau: a 24/7 AI assistant for customers via the OneService APP, handling common queries on policy status and renewals, and an AI knowledge retrieval tool for financial advisors. Both are built on the DeepSeek V3 large model, emphasizing natural language understanding and compliance with local regulations. The design principle is to have AI handle straightforward issues instantly while routing complex matters to humans during office hours, enabling human-machine collaboration.
The impetus for this innovation is underscored by the insurer's operational scale. In 2025, China Life Overseas reported robust performance, with total premium income reaching HKD 49 billion and value-based new business premiums surging 63% year-on-year to HKD 10.5 billion. Comprehensive investment income totaled HKD 22.15 billion, with an annualized rate of about 6%. As policy volumes grow, so does consultation demand, making manual responses slow, costly, and error-prone. The insurer's stable operational results provide the foundation for exploring such technological advancements.
Comments