2023 is considered the first year of the big model. ChatGPT surpassed 100 million monthly users at the beginning of the year and quickly became a high-profile phenomenon. At the same time, it also kicked off commercial competition centered on big models.

Throughout the past, every wave of technology was not just a competition for technology itself. The general model represented by Chatgpt has exploded the market. Next, the common proposition faced by AI-related companies around the world is how to achieve the true implementation of technology in human production and life scenarios. At this point, the vertical industry model has naturally become a new outlet for the market.

In this context, Gelonghui discovered that in the vertical field of AI+ big consumption, Xuanwu Cloud (2392.HK) is not only at the forefront of the industry at the technical level, but AIGC related product applications have also enabled consumer goods customers to land. According to data, Xuanwuyun set up an AI laboratory in 2016, and is currently the only CRM manufacturer with AI full-stack self-developed technology in the industry. In 2022, Xuanwu Cloud's AI-related revenue increased 73.8% year over year; in 2023, while launching a multi-modal model for the vertical consumer sector codenamed “Xuantao,” Xuanwu Cloud also reached a strategic cooperation with HUAWEI CLOUD, and the two sides will jointly create a big model for consumer goods.

As the largest smart CRM service provider in China, Xuanwu Cloud has always adhered to the “technology+business” two-wheel drive strategy. Compared to traditional CRM, Xuanwu Cloud's smart CRM based on AI, DI, and APaaS technologies can not only reduce costs and increase efficiency for customers, but also help customers achieve sustainable sales growth through digital intelligence technology.

Recently, Chen Yonghui, Chairman and CEO of Xuanwuyun, was interviewed by Gelonghui. In the interview, Chen Yonghui explained in detail how the company achieved collaborative technology+business development, and how to follow the second growth curve of “AI+ big consumption” in the big model era.

Picture: Chen Yonghui, Chairman and CEO of Xuanwuyun

The following is a transcript of Gelonghui's exclusive interview with Chen Yonghui:

Focus on the big model application layer to help consumer goods companies achieve terminal sales growth

Gelonghui: The current trend of big models moving towards various industries is obvious. The company took the lead in releasing a multi-modal big model perpendicular to the big consumer industry. Please briefly introduce the company's capabilities and applications in the field of AI and large models?

Chen Yonghui:As for the development of AI technology, we have always been researching and deploying, so in terms of technology, we have a very good accumulation, which is an important foundation for us to quickly launch a multi-modal big model of big consumption. The reason for focusing on the field of big consumption is that we have been deeply involved in the consumer goods industry for many years, focusing on providing solutions for various business scenarios of consumer goods companies, and have accumulated rich industry data as a result. This allows us to pre-train the model with sufficient amount of data, and then build the AIGC application layer on top of the big model base.

If it is the ability to just look at large models, our “Xuantao” big model can show advantages such as generality and a leap in modeling efficiency, from perception to multi-modal perception, intelligent understanding of business knowledge, and intelligent integration of large and small models. However, what we value more is to combine the customer's business scenario and use large model capabilities to help customers better achieve cost reduction, efficiency and sales growth.

Currently, the company has two main application ideas: first, make full use of the “Xuantao” multi-modal model to upgrade and update the original business scenarios; second, explore more application scenarios in large consumer fields.

Based on the big model, we have also released AIGC applications, such as display commanders, visit and summary assistants, etc. Take the application capabilities of display commanders as an example. Through multi-modality and AIGC capabilities, it can help FMCG brand terminals quickly inspect stores, automatically detect and generate sales value for terminal stores, and also generate suggestions for current product display improvements and product quantity increases or decreases, thus helping brands build “perfect stores.” The data accumulated during this process can also later provide data support for brands' terminal fee investment and marketing strategies.

Gelonghui: Recently, the company also cooperated with HUAWEI CLOUD to create a major consumer product model. What impact will this have on the company's future?

Chen Yonghui:At present, the company has made some progress in the field of big models. Our cooperation with HUAWEI CLOUD's “Pangu Big Model” will first reinforce and upgrade our big models in terms of technical capabilities, which will help the company build products related to big models in the future. Second, we are already planning to conduct scenario function tests with customers using the company's AI products. In the future, we will work with HUAWEI CLOUD to explore more product applications and market development in large consumer business scenarios, which will also benefit our business development in the large consumer sector.

Additionally, we are old friends with HUAWEI CLOUD. Currently, we have developed multi-dimensional cooperation in other business areas. For companies, joining the HUAWEI CLOUD open ecosystem and growing with industry partners is more important than simply improving technical capabilities.

Gelonghui: With the advent of the big model era, all AI vendors are facing huge challenges in acquiring and cultivating talents, especially in terms of algorithms. How does Xuanwu Cloud cope with this challenge?

Chen Yonghui:First, we have built an excellent internal technical team, and Xuanwu Cloud's AI lab has been dedicated to cultivating talents in algorithms since its establishment in 2016. Internal training is one of Xuanwuyun's core talent team building methods. At present, we have an AI team led by top scientists, which is the cornerstone of the company's future AI development.

From another perspective, the company focuses more on R&D and innovation at the vertical application layer of the large model rather than on the development of the underlying large model. The company previously used some open source models, such as Transformers. At the basic model level, we will not compete with similar companies such as Open AI. Mathematicians dealing with low-level math problems are different from those of developers who build applications on models, and competing across borders is a challenging task.

Therefore, we will also appropriately seek technical support from major external manufacturers, which is why we chose to cooperate deeply with HUAWEI CLOUD. In the HUAWEI CLOUD ecosystem, they will provide some technical support to partners, especially at the bottom of the big model. In this way, we can focus more energy on the AIGC application layer to develop more application scenarios for large consumer industries.

Adhere to technological self-research and innovation, and win customer trust with product value

Gelonghui: Based on the business value provided by Xuanwu Cloud to customers, how do you think the company can continue to guarantee customer stickiness?

Chen Yonghui:Xuanwu Cloud's smart CRM solution is not only a core component of the customer's business system, but is also essential for customer operation and revenue generation. It covers all customer business processes. Once replaced, customers will need to readjust and change usage habits.

Second, all of the customer's data and reports have been settled into Xuanwu Cloud's smart CRM solution. Replacing this system is expensive, probably much higher than purchasing a new system. Furthermore, compared to other systems such as OA or ERP, CRM involves a wider range of people. In summary, smart CRM is hard to replace unless disruptive and innovative technology comes along.

Glonghui: For a company, constantly acquiring new customers is the key to continuous growth. Does the company have any “winning strategies” in this regard?

Chen Yonghui:Acquiring customers is not an easy task for any TOB AI company or SaaS company. Currently, most of the main customers served by TOB companies are usually medium to large enterprises, and the decision-making cycle of medium and large enterprises is longer than that of small enterprises. Additionally, customers may require business changes when introducing new business systems, which determines the longer customer acquisition cycle for individual customers.

Changing this situation isn't easy, so we're actively looking for more ways. On the one hand, we insist on continuous innovation in technology and products. Like the big model I mentioned earlier, AIGC, and our DI, these technologies enable our smart CRM to create a “quantitative growth” model for consumer goods companies. Under this model, our products can not only reduce costs and increase efficiency for customers, but also help customers achieve continuous growth in terminal sales. This cannot be achieved by traditional CRM, and it is also more attractive to customers.

At the same time, the company is also trying to attract customers by conveying a better value system, including sharing successful experiences by creating benchmark cases and establishing 10 billion clubs, and in view of the long decision-making cycle of individual customers common in the industry. We are gradually standardizing our overall solutions to cover more small and medium-sized enterprises; what cannot be overlooked is that we are also considering whether we can expand our business through ecological cooperation with companies such as Huayun and use their channel advantages to help the company open up the market faster.

Glonghui: Are there any trial pay-for-results strategies so that customers can experience the benefits that the product may bring first?

Chen Yonghui:We're trying to use experiential sales, but I think it will only be possible if we provide some small to medium customer experiences based on standardizing product solutions.

For example, AIoT smart freezers, which are currently an extended scenario application of our sales terminals in the consumer sector, are actually highly standardized AI products and services, which enable customers to quickly obtain experiences. Customers can try standardized smart freezer products. For example, by testing 100 or 200 devices in some cities, customers can evaluate the actual effect of the product and the degree of compatibility with their business before making a decision. As a result, customers can confidently decide whether to promote it on a larger scale.

Smart freezer products aim to solve the problems of some large dairy and beverage companies in terminal retail freezer management. In the past, these companies placed their own freezers in stores to sell their products, but the products in everyday freezers may be replaced by other companies, making it difficult to effectively manage the display, thus affecting terminal sales. The smart freezer is equipped with a camera to photograph the display situation and record each sales situation, thereby monitoring inventory conditions in real time, ensuring product display optimization, and further increasing terminal sales.

In addition, our standardized products also include smart store expansion services to help brand customers find suitable offline terminals. Customers can try within a region for a period of time before deciding whether to pay. Overall, the trial model is effective for customer acquisition, but this model is more suitable for standardized products.

Gelonghui: What do you think is the company's core competitive advantage compared to competitors in the big consumer sector?

Chen Yonghui:Like I mentioned above, technology and innovation.

The company has a high technology self-development rate. Because our team is from technology, we pay more attention to self-research technology. We invest relatively heavily in R&D and have leading innovation in our segments.

Compared to face recognition, product recognition in large consumer sectors faces more complex challenges. Due to the variety of packaging specifications of consumer goods and rapid updates, the same type of product may present various packaging forms depending on the manufacturer, region, or promotional activities. Fast and accurate identification of terminal store SKUs actually tests the recognition technology ability of AI manufacturers. Therefore, we have been continuously upgrading our algorithms and models at the technical level. Currently, the overall recognition rate of our AI technology has exceeded 92%, ranking at the leading level in China.

Furthermore, our technological innovation has always been closely integrated into scenarios, usually deeply integrated with customer needs. For example, Xuanwu Cloud establishes data services to help customers find suitable offline terminals and help them better manage their business through data analysis. The company's AI platform and DI platform are all built based on customer business needs and technological changes. In the future, we will follow customer needs more closely, grasp technological changes, and continue to build technical capabilities.

Keep up with changes in market demand and take more measures to drive business growth

Gelonghui: Do you think there are any changes in the marketing needs of consumer goods companies now? Does this change have an impact on the company's business development?

Chen Yonghui:One obvious characteristic of today's consumer goods companies in terms of spending is that they are cautious and only spend effective money. This trend is probably more pronounced than before.

In the past, some businesses might only need a lot of advertising to achieve phenomenal success before expanding more dealers. However, this kind of growth has become more difficult in the post-pandemic era, because the current stock market, channels are already very saturated, and market competition has become more intense. As a result, the importance of digital intelligence transformation has been highlighted: smart CRM can not only help enterprises control costs and improve efficiency, but also help enterprises truly increase terminal sales by finding the perfect store and improving display. Obviously, this kind of product solution, which can increase their own performance, will make consumer goods companies more interested in us.

Gelonghui: We have noticed that in the 2023 interim report, the company proposed a second growth curve focusing on driving AI+ consumption. How was this strategy considered?

Chen Yonghui:First, consumer goods retail has a very large market space, and in this field, we will have plenty of room for imagination. Second, the company has been deeply involved in the consumer goods sector for more than ten years, especially FMCG products. Our market share of the top 100 FMCG companies in various segments is over 20%, and we already have a certain market leadership. However, when we are deeply involved in the industry, we discovered that in consumer goods retail, “people,” “goods,” and “markets” are the main focus of consumer goods companies, and there are many potential opportunities to use AI to reshape the value chain, so there is a large market space. In summary, whether from the characteristics of the consumer industry or from the path of business development, it is a natural process for us to introduce AI into the consumer sector.

Gelonghui: What changes will the second growth curve have on the current composition of the company's business revenue?

Chen Yonghui:In terms of revenue, the AI+ consumer business is a high-margin business. By standardizing our products and solutions, we plan to further increase the gross margin level of this business segment while reducing implementation and delivery costs. This has had a certain effect on the steady growth of the company's overall gross margin.

In addition, there may also be an increase in the way customers charge. Currently, Xuanwu Cloud mainly charges fees through software licenses, etc. And after having a large model, Xuanwu Cloud may charge the customer in terms of the number of calls made to the large model.

Gelonghui: How will the company acquire incremental customers in the future?

Chen Yonghui:Our current plan is to achieve a breakthrough in two areas:

1. Continue to focus on the big consumer sector. In the past, Xuanwuyun has achieved remarkable results in the FMCG field, covering many industry segments, including beverages, food, cereals and oils, liquor, beer, cosmetics, health products, etc. These industry segments themselves are highly scalable, as they share similar channel models, and market potential can be further expanded.

2. Achieve breakthroughs in customer groups. In the past, Xuanwu Cloud mainly chose medium to large customers because they had strong purchasing power, ability to change, and financial strength. Now, as the business model matures, Xuanwu Cloud will also focus on emerging brands and small to medium brands through standardized products. Xuanwu Cloud can directly provide mature standardized products to this part of the customer base.

In addition, we are also exploring the possibility of providing micro-consulting services, such as providing advice on new categories of team building, channel incentives, and data system establishment. Through digital product systems and micro-consulting, Xuanwu Cloud can better penetrate new categories and provide services to more emerging customers. This service method has more extensive penetration capabilities and helps expand Xuanwu Cloud's growth space.