With the support of AI large model technology, Tencent Qidian’s intelligent customer service and marketing analysis capabilities have been upgraded.

On September 7, at the 2023 Tencent Global Digital Ecosystem Conference, Tencent Qidian announced a number of latest developments: Based on AI large model capabilities, it will improve the efficiency and user experience of Tencent Qidian intelligent customer service management and open Tencent to customers. Qidian Analysis·AI assistant internal beta application helps corporate customers create new growth engines through "conversational analysis".

Wu Yongjian, vice president of Tencent Cloud, head of Tencent Cloud Intelligence R&D, and head of Tencent Qidian R&D, said: "The large-scale industry model built by learning a large amount of professional knowledge and strong reasoning capabilities has great performance in specific fields. and more adaptable. In the future, we will continue to improve the production efficiency of intelligent applications such as intelligent analysis and intelligent customer service based on the continuously upgraded industry model, and bring more diversified, plug-and-play products and services to the market. ”

Large model intelligent customer service improves the efficiency of knowledge construction and increases the resolution rate of complex problems by 30%

Traditional intelligent customer service relies on knowledge base operations to provide answers to customer inquiries. On the one hand, the huge amount of knowledge maintenance brings a burden to corporate customer service operations. On the other hand, limited by limited knowledge boundaries, traditional intelligent customer service cannot answer questions that are not in the knowledge base and provide anthropomorphic interactive experience, facing the " Unintelligent” challenge.

In June this year, focusing on customer service scenarios, Tencent Qidian launched a new generation of intelligent customer service, introducing large model capabilities and improving the three levels of dialogue interaction, manual assistance and knowledge construction. In the latest product upgrade, Tencent Qidian Intelligent Customer Service further upgraded the management-end configuration capabilities, improved the efficiency of knowledge construction, optimized the user-end experience, and completed multiple rounds of business processing through anthropomorphic and chat-style methods, making the intelligent customer service closer to "real-person customer service" ".

At the level of knowledge construction, the new generation of intelligent customer service improves the reading and understanding effect of complex documents through the closed loop of "import-verification-tuning" of enterprise-specific knowledge, further optimizes cold start efficiency, and optimizes cold start from 7 days To the minute/hour level, the startup accuracy rate is increased by 30%, reducing operation and maintenance costs. For example, on the basis of text documents, it has added access to complex documents such as mixed graphics and text, multi-column typesetting, etc., and can generate vectors through semantic slicing, and use a highly available and scalable vector database solution to enrich question and answer combinations and content; By quickly generating question and answer pairs and tracing back the original text of the question and answer, the efficiency of question and answer verification is improved; by providing dialogue testing and operation tools, operators can quickly implement tuning and improve the accuracy of question and answer.

At the level of dialogue interaction, large model iteration vector database and search engine capabilities can intelligently generate anthropomorphic answers, allowing intelligent customer service to respond to more complex needs. For example, traditional customer service cannot support complex requirements such as "explanation of operation steps", but large-model intelligent customer service can complete contextual information and perform question and answer reasoning based on the enterprise knowledge base, increasing the resolution rate of complex problems by 30%.

In terms of user experience, the support of large models also makes the intelligent customer service experience closer to "real customer service". For example, the new generation of intelligent customer service can more accurately identify user emotions and provide timely comfort; it can distinguish task-based, knowledge-based, and chat-based topics, and use the tone and method of chat to solve task needs. Even if the user inserts a "chat" topic during the business processing process, the task will not be interrupted, which improves the business processing success rate of intelligent customer service.

Tencent Qidian Marketing and Analysis 6.0 major upgrade, AI assistant open for internal testing application

The rapidly changing marketing environment continues to pose new topics to enterprises, such as realizing "all-area marketing" covering all channels and the entire journey, implementing refined and customized "personalized marketing", and providing data-driven and intelligent "Digital Intelligence Marketing", "Integrated Marketing" that connects new retail online and offline, etc.

In order to better solve the pain points and challenges of corporate marketing, Tencent Qidian has newly upgraded and released the "Qidian Marketing and Analysis 6.0 Product Solution" to help enterprises better respond to market changes and achieve marketing goals.

According to reports, the newly upgraded Qidian Marketing and Analysis 6.0 has three distinctive product capabilities: integration, scenario-based and intelligent.

Among them, "integrated" product capabilities help enterprises shorten the entire process of data access, preparation, analysis, application and verification, and quickly transform data value into business value; focusing on "scenario-based", with the help of big data and AI technology, Integrate data from multiple brand touch points to accurately identify user identities and journey stages, provide differentiated marketing content and strategies in different scenarios, and improve customer experience and brand marketing conversion effects.

In the process of fully embracing the large AI model, the newly upgraded Tencent Qidian Marketing and Analysis 6.0 has also built unique "intelligent" product capabilities. On the one hand, based on the large analysis model, Tencent Qidian Analysis AI Assistant has been upgraded and launched, covering four scenarios: conversational analysis, auxiliary data configuration, extraction of intelligent conclusions, and one-click report generation, allowing everyone to have simple natural language conversations. , obtain professional data insight results, and achieve a leap-forward improvement in analysis output efficiency. At present, Tencent Qidian Analysis·AI Assistant 1.0 has been implemented and verified by multiple leading customers in retail and other industries, and has officially opened internal testing applications to all customers.

On the other hand, with the help of AIGC-oriented large model capabilities such as Wenshengwen, Wenshengtu, and content stylization, Tencent Qidian Marketing will also upgrade its AI assistant capabilities to help companies improve the efficiency of personalized marketing content production and obtain better results. conversion effect and customer reputation.

Currently, Tencent Qidian has served more than 80 industries, such as industry, retail, finance, cultural tourism, real estate and education, connecting more than 750 million enterprises and users. In the future, Tencent Qidian will continue to use the powerful capabilities of the latest technologies such as large models to continue to promote the intelligent upgrade of full-process solutions for corporate marketing, sales, operations and services, and achieve common growth with customers.

The above is the detailed content of With the support of AI large model technology, Tencent Qidian’s intelligent customer service and marketing analysis capabilities have been upgraded.. For more information, please follow other related articles on the PHP Chinese website!