Driven by the wave of digitalization, the insurance industry is ushering in an efficiency revolution led by AI large model technology. Recently, Nuanwa Technology said that with its business accumulation in the insurance field, the company has successfully applied AI large model technology to insurance intelligent marketing and customer business scenarios, achieved a significant improvement in the efficiency of private domain operations, and is committed to injecting new vitality into the insurance industry.
It is understood that based on the iFLYTEK base model Xinghuo 3.5, combined with the company's many years of vertical application business data in the insurance industry, through the AIGC core technologies such as large model supervision and fine-tuning (SFT) and knowledge distillation, Nuanwa Technology has successfully developed the insurance vertical model - Alamos V1.0.
In the private domain scenario, Alamos transforms the traditional operation SOP into a full-life cycle customer management agent driven by a large model, which can realize functions such as active user activation, accurate identification of user intent, agent reply speech generation, and sales rhythm orchestration, and improve user satisfaction and reduce the rate of complaints and deletion while increasing the GMV of the private domain, and greatly improve the productivity and efficiency of customer management agents.
Alamos V1.0 has been in production for nearly half a year and continues to iterate. According to statistics, with the launch of Alamos, the overall production capacity of Nuanwa's private domain has increased by 20%, of which the single-node capacity conversion rate and GMV of the main process of the user life cycle have doubled, and the accuracy rate of Alamos in user intent recognition has reached more than 90%, and the acceptance rate of activated voice generation agents has reached 92%, which has been widely praised by front-line operation and sales personnel.
Scenario examples:
First of all, Alamos can select users who can be activated by marketing according to operational goals, user journeys and user portraits, actively generate words and tasks, and assist agents to judge the intention of user replies and give suggested responses, which can complete multiple rounds of dialogue and provide users with more personalized services.
Scenario (1): Voice activation invitation of enterprise and micro customers. Differentiated guidance and activation are carried out according to the reply of the incoming customer and the specific reply content.
[T+1 agent active follow-up activation]
[T+2 Follow up again to guide the voice]
Alamos will carry out customer operation and service throughout the life cycle of users, and will actively generate tasks and content to activate users based on changes in customer transaction status, physical status and claim status, so as to achieve user services throughout the life cycle.
Scenario (2): The insurance policy of the user in the agent enterprise micro-enterprise is not renewed, and the Alamos customer management agent actively generates tasks and words to remind the agent to complete the renewal recall.
【Renewal Service】
In the private domain marketing scenario, the Alamos V1.0 model will rhythmically choreograph the original T+X automated marketing reach strategy, accurately identify and reduce the disturbance of customers with negative emotions, and dynamically adjust the reach time based on the dialogue context and user portrait.
Scenario (3): Optimization of enterprise and micro automation marketing strategy.
In addition to the enterprise WeChat scenario, Alamos also has a prominent application in the telemarketing scenario, where agents can realize real-time objection identification when making outbound calls, give real-time reasons for customers not to close the deal and recommend ways to resolve objections, and automatically generate call summaries, redial times and priority suggestions after completing the call.
Scenario (4): In the online telemarketing scenario, customer objection points are identified and recommended for speech.
[Outbound Call Workbench: Suggestion response dialogue after objection recognition through real-time voice translation]
In the design of Alamos, Nuan Wow Technology has adhered to several principles from the beginning. On the one hand, Alamos is not a marketing agent, not only pursues the maximization of sales, does not sacrifice customer satisfaction, and can serve customers according to the rhythm set by the operation. On the other hand, Alamos is not a large customer service model, it is aimed at customer management, and takes the initiative to initiate services to customers instead of only responding to customer incoming lines. Finally, Alamos aims to effectively increase sales efficiency, improve service and sales staff efficiency under the premise of increasing compliance and user satisfaction, and achieve overall efficiency.
At present, after Nuanwa Technology applied AI large model technology to insurance intelligent marketing, the online telemarketing business has achieved remarkable results, the overall sales capacity has increased by more than 20%, the average daily number of users has also increased significantly, and negative indicators such as the deletion rate of customers at a single point have also been effectively reduced. These data fully prove the application effect of Nuanwa Technology's AI large model technology in the operation of private domain customers.
Nuanwa's practice in the application of large models has opened up new imagination for intelligent marketing in the insurance industry. In the future, Nuanwa Technology will continue to deepen the AI large model technology, promote the intelligent upgrading of insurance business, and empower insurance companies to provide more high-quality and efficient insurance services for more users.