Ant Group’s AIGC Application ‘Zhi Xiaobao’ Has Over 10 Million New Users

The AIGC track in China has entered the stage of competing in industrial applications. On April 11th, Ant Group‘s AI Financial Assistant ‘Zhi Xiaobao’ 2.0 version is already in the external testing phase. Zhi Xiaobao is positioned for financial and insurance professional knowledge Q&A, while also providing tools such as market analysis, portfolio analysis, intelligent underwriting, and claims processing. Currently, within the Alipay app’s Ant Wealth and Ant Insurance channels, users can access Zhi Xiaobao services which support text and voice Q&A. The interactive experience will continue to be rapidly iterated upon. It is understood that the new version of Zhi Xiaobao has been experienced by 12 million users with terms like “market trends” becoming the most popular questions asked by users in the first quarter.

Compared with ERNIE Bot, Kimi and other domestic chatbots, Ant Group’s Zhi Xiaobao is positioned to answer professional questions in the financial vertical fields such as funds and insurance, providing reference analysis and reminders. Testing found that Zhi Xiaobao can provide practical tools based on users’ financial and insurance product holdings, including financial holding analysis, market trend interpretation, as well as intelligent underwriting and intelligent claims processing for insurance. For example, when asking about a certain fund, Zhi Xiaobao can provide comprehensive product data and also alert about risks like scale and changes in fund managers.

The data released by Ant Fortune shows that users who have been deeply served by Zhi Xiaobao have a 5% higher level of reasonable asset allocation compared to those who have not been served, and the proportion of users engaging in frequent trading has decreased by 60%.

In the insurance field, Zhi Xiaobao provides “smart underwriting” capabilities, which can assess whether users can be insured based on their health conditions. The “smart claims” function can achieve fully automated claims review for some simple outpatient insurance claims.

According to the introduction, the financial knowledge data storage of Zhi Xiaobao 2.0 has reached a scale of billions and uses a “multi-intelligent agent collaboration” model to simulate human experts in different aspects, providing opinions and accessing professional service tools. Currently, Zhi Xiaobao’s ability for financial event analysis and reasoning has reached the average level of human industry experts.

Previously, domestic AIGC has started exploring vertical fields such as healthcare and transportation. Zhi Xiaobao is the first AI application in the financial field to achieve large-scale customer interaction. Zhang Hua, Vice Dean of the Finance Research Institute at Zhejiang University, analyzed that “these vertical fields have high professional thresholds and compliance requirements. AIGC needs stronger knowledge data, logical abilities, and technical capabilities in these areas. Once successfully applied, it will significantly reduce service barriers in these industries and enhance user experience.

Since the beginning of this year, A-shares have been fluctuating and warming up. There have been continuous news about lowering house prices and interest rates, causing investors to be cautious and seek asset preservation. Many users have consulted Zhi Xiaobao.

SEE ALSO: Ant Group’s Intelligent R＆D Platform CodeFuse Launches the ‘Graphical Code’ Feature

The question-and-answer data for the first quarter of Zhi Xiaobao shows that the two most frequently asked types of financial questions by users are “market analysis” and “weekly returns.” Questions like “Which health insurance is suitable for pre-existing conditions?” and “When can I receive savings insurance benefits?” have become the hottest insurance issues. When providing professional answers, Zhi Xiaobao also warns about market risks and reminds users to make rational choices.

The data also shows that since 2024, the new version of Zhi Xiaobao has answered a total of 8.45 million financial and insurance questions manually entered by users, with an adoption rate for corresponding answers increasing by around 20% compared to the previous version.