Application Case: Business application and experience empowerment of artificial intelligence

How can big data as a factor of production, artificial intelligence (AI) as productivity, and experience focus on people themselves?

During the 2023 User Friendly International User Experience Conference held in Shenzhen, Mr. Guo Xiaobo, Senior Vice President of Zhongyan Technology, at the Data Intelligence Summit, with the theme of "Digital Intelligence Changing Experience and Welcoming a New Era of Business", in-depth insights into digital intelligence in business prospects and experience empowerment, and revealed the business value of generative AI in enterprise R&D, production, marketing, customer service and other fields through rich cases.

1. The business prospect of digital intelligence

1. From decision-making AI to generative AI

From the three dimensions of computational intelligence, perceptual intelligence and cognitive intelligence, Mr. Guo enumerated some typical events in the development process of artificial intelligence. And big data also brings big models. The number of parameters of the GPT series models has grown rapidly, from the initial 117M to the GPT-4, which is reported to have reached 176T, GPT-4 has superb performance in general knowledge and logic. The era of big data has played a huge role in promoting the development of models.

2. Core values and business applications.

Through a series of enterprise digitalization cases, Mr. Guo demonstrated the commercial applications of generative AI in different fields. Digital intelligence has brought great business potential.

R&D – Using the CALA fashion design platform as an example (see below), generative AI provides designers with rapid creative support, significantly lowering the barrier to entry for new designers.

Take the structural design of the Goddard Space Center as an example ( ) by comparing the difference between the effects of artificial design and AI design, and feel the amazing efficiency and effectiveness of AI when designing complex structures.

Production – Google's RT-1 is used as an example to demonstrate the efficient learning and processing capabilities of generative AI robots for multitasking. In addition, through the application case in the field of industrial quality inspection ( ) using the generation of defects** for model training, the problem that it is difficult for traditional models to obtain diverse data is solved, and the accuracy of defect detection is improved.

*——Taking Wumart supermarket as an example, through multi-dimensional data analysis, compared with the traditional model, the accuracy of the first product has been improved, the out-of-stock rate of goods and the number of inventory turnover days have been significantly reduced, and the chain management has been greatly optimized.

Marketing - Take the e-commerce platform Zalando as an example ( ) Customers can transfer clothing colors or postures to different models, which realizes the rapid production of marketing ** and content, and reduces repetitive work.

In addition, the application of virtual anchors ( ) provides 24-hour uninterrupted product recommendations and services, and there is no risk of collapse of the personality.

Customer Service – Intercom's FIN is used as an example to illustrate the continued evolution of AI customer service bots in the customer service space.

2. Digital intelligence empowers the experience

The application of AI technology to digital and intelligent customer experience management is a hot topic in customer experience management. As the first service provider in China to integrate the closed-loop of experience indicators questionnaire design, data collection and processing, intelligent calculation and analysis, visual BI presentation, and driving timely action, Bestco has solved the problem of "hearing clearly and seeing accurately" through data collection and intelligent analysis, and solved the problem of "being able to do and use it steadily" through landing actions and experience operations, and has collected more than 400 million experience data for more than 20,000 outlets in more than 30 industries, and promoted the first time experience repair and action recovery.

In 2023, Baseworld was also honored to be ranked No. 1 in market share in this segment by IDC, and ranked No. 1 in "Journey Analysis and Orchestration Tools" and "Experience Management Tools" by UXACN.

Although Baseline introduced AI capabilities early on, this year's large-scale model boom has enabled Bestco to carry out a comprehensive upgrade of product capabilities, upgrading the AI capabilities that were originally only used for analysis to the AIGC large model capabilities of the whole site, and building the basic services of Xiaobei AI, including intelligent generation of questionnaires and delivery suggestions, optimization of opinion extraction and sentiment classification in text analysis, etc., to improve the efficiency and effectiveness of use.

1.Intelligent questionnaire design and optimization.

Designing an effective questionnaire is not an easy task, for example:

· Clear purpose and clear logic.

Pay attention to the accuracy, consistency, and sensitivity of the issue.

Set up your answer options wisely.

Avoid misleading and distracting answers.

Avoid psychological discomfort for the respondent.

Clever authenticity testing.

2.Advantages of large models: efficiency improvement and multi-language processing.

In terms of text analysis, compared with the traditional small model, after docking with the large model, Basestar has demonstrated excellent performance in semantic understanding, judgment, classification, etc., eliminating a large number of annotation, training and multilingual translation, improving the output efficiency by more than 80%, increasing the accuracy by about 20%, and removing the obstacles to multilingual application in globalization.

3.BX Research & BX Stats: New decision-making models and algorithmic support.

In terms of data analysis, Baseline has introduced a new decision-making model and algorithm to support BX Research, which integrates the most commonly used decision-making models, and integrates questionnaires, sample services, algorithms and reports to help users generate report results with one click in a short period of time, without having to learn the complex algorithm model behind it. At the same time, BX Stats has connected with the SPSSPRO algorithm supermarket, which can provide 360+ algorithm models, and has provided professional, flexible and efficient data analysis for more than 1 million analysts.

4.BX Club: Build the main position of brand experience operation and refine experience operation.

Basestar has also launched the BX Club, a brand experience officer, to help brands establish an independent and controllable private domain operation position, and support various special activities for experience operation, such as experience officer recruitment, new product trial and retention test, intra-city activities, fan community and mystery guest. Brands can quickly access the Mini Program template of Baseline and independently define recruitment requirements, growth system and rights and interests. It not only provides decision-making support for the brand, but also improves user stickiness and realizes the precipitation of private domain assets.

Through this round of innovation, Baseline not only provides more explicit capabilities in design, collection, analysis, operation, and action, but also continuously improves its back-end capabilities. Through unremitting efforts, we look forward to pushing experience management into a new stage of rapid development, and working with customers to create a better experience for consumers.