Big data industry innovation service media
——Focus on data·Changing business
After several years of “digital transformation”, the potential of “user touchpoints” in corporate customer contact scenarios has been quickly unleashed. With the help of mature enterprise customer contact solutions, it can be more efficiently integrated into the enterprise's "all-domain operations" system, generating greater value in terms of public domain traffic conversion and private domain traffic awakening.
In this industry context, the market demand for enterprise customer contact solutions is increasing day by day. The "2023 China Intelligent Customer Service Market Report" released by Sullivan, a global growth consulting agency, pointed out that the market size is expected to grow to 18.13 billion yuan by 2027. As an AI-driven full-cycle customer contact solution SaaS cloud platform, Hong Kong-listed company Tianrun Cloud (02167.HK) released its unaudited interim results for 2023, with revenue of 212.9 million yuan, a year-on-year increase of 10.5%.
Regarding business growth, the financial report claimed that "mainly due to the increase in SaaS solution revenue generated from premium customer subscriptions, the revenue contribution of Beijing Yizhang Yunfeng Technology Co., Ltd. (has been acquired by the Group)" and "the rapid expansion of AI-related product portfolios" ".
The "AI related product portfolio" mentioned therein is the Wei Teng large language model launched in March 2023 (hereinafter the large language model will be referred to as the "large model"). After several months of “battle of hundreds of models”, there is no lack of discussion on the topic of “the last mile of commercialization of large models” in the industry. As for the SaaS cloud platform positioned as an AI-driven full-cycle customer contact solution, Tianrun Cloud can be used as one of the cases.
Highlights of the financial report: SaaS revenue increased by 18.5% year-on-year despite the adverse effects of the education industry.
According to the financial report, Tianrun Cloud can provide a variety of cloud-native customer contact solutions - communication solutions that enable enterprises to interact with customers through multiple channels, and conduct sales, marketing, customer service and other businesses in an intelligent manner. functions. At the same time, the company has "established an extensive, high-quality and loyal customer base" involving different industries such as technology, insurance, automobiles, education, health care, banking, manufacturing and logistics. From January to June 2023, Tianrun Cloud's revenue totaled 212.9 million yuan, the company's gross profit was 102 million yuan, a year-on-year increase of 11.1%, and the gross profit margin increased from 47.7% to 47.9%.
According to the prospectus submitted by Tianrun Cloud to the Hong Kong Stock Exchange in June 2021, its main customers are concentrated in technology, education and other industries. Among them, customers in the education industry account for about 30% of Tianrunyun’s revenue. Therefore, due to the strong regulatory environment in the education industry, its revenue has declined after growing for many years. In this regard, it also made corresponding explanations in the prospectus at the time, saying that "recent regulatory changes will lead to a significant reduction in revenue from customers in the education industry and have a significant adverse impact on its business, operating results and financial conditions."
As for this revenue growth, in the semi-annual report released this time, it was attributed to "the increase in SaaS solution revenue generated from high-quality customer subscriptions and the revenue contribution of Beijing Yizhang Yunfeng Technology Co., Ltd. (has been acquired by the Group)" ".
After completing the acquisition of Beijing Yizhang Yunfeng Technology Co., Ltd., the company's business brand Huanxin has successfully completed its business reorganization. Its customer service cloud business was delivered to Tianrunyun and continues to operate as a wholly-owned subsidiary of Tianrunyun. This business occupies a leading position in China's SaaS customer service market, covering insurance, banking, automobiles, logistics, travel and many other fields, and has served more than 10,000 corporate customers.
In the first half of 2023, Tianrun Cloud had a total of 4,222 customers, a year-on-year increase of 58.8%. Its SaaS customer retention rate and the net retention rate of all SaaS customers based on amount were 74.7% and 105.4% respectively. For comparison, in the first half of 2022 They were 74.7% and 80.7% respectively.
Among them, SaaS business revenue was 198.6 million yuan, a year-on-year increase of 18.5%. In this business, the total revenue of intelligent contact center solutions is 170.5 million yuan, the revenue of remote agent solutions is 21.4 million yuan, and the revenue of ContactBot solutions is 066.8 million yuan.
The financial report also mentioned that "technology is the core foundation of our solutions." In the first half of 2023, Tianrun Cloud "continues to enhance our technological leadership position through product and technological innovation", and the R&D team increased by 5.9% from 253 (as of June 30, 2022) to 268 (as of June 30, 2023) day), accounting for 45.8% of its total employees.
In the first half of 2023, Tianrun Cloud products will maintain rapid iteration, "releasing new versions approximately every week". At the same time, the system uptime will reach 99.99% (calculated based on the percentage of time customers use and operate it), and the platform has been in operation for more than 48 months. There has been no global system failure during the period.
Promote Weiteng language model and integrate it into customer contact scenario solutions
According to the financial report, Tianrun Cloud, as an AI-driven full-cycle customer contact solution SaaS cloud platform, independently developed and realized the deep integration of "AI, cloud and communication" technologies. In March 2023, the company launched the Weiteng language model (platform) to "integrate intelligence into the process of customer contact solutions", focusing on corporate knowledge engineering and promoting corporate growth.
In the past August, Tianrun Cloud also launched the Weiteng Language Model (Platform) 2.0, saying that it "significantly improved the competitiveness of SaaS products and became a key driver of business growth."
According to public information, the Weiteng large language model platform is composed of mainstream open source basic models and mainstream general large models. By tuning these large models, the models are matched to relevant customer scenarios. At the same time, it also integrates plug-in market, prompt word factory, open API, etc.
In addition, the Weiteng language model platform also takes enterprise knowledge management as a breakthrough, "by consolidating the enterprise knowledge base, intelligently empowering various processes such as marketing, sales and services", providing efficient assistance for the entire process of customer contact, using In business scenarios such as customer service reception, session analysis and extraction, automatic form filling, etc. At the same time, the Weiteng big language model "can allow enterprises to more accurately understand customer needs and expectations through in-depth analysis and insights of large amounts of data", thereby optimizing their marketing strategies, product design and customer services, and achieving "knowledge-driven enterprise high-speed increase".
Exploration of typical commercialization practices of large models: used in the direction of enterprise knowledge engineering
In its financial reports and public information, Tianrunyun claimed that its Weiteng language model focuses on "enterprise knowledge engineering." After years of development, knowledge engineering construction has been valued by more and more enterprises, and its core purpose is to help enterprises create personalized knowledge service systems.
During the release process of Weiteng Language Model (Platform) 2.0, Tianrunyun believed that enterprise knowledge mainly includes three categories: the first is the original enterprise knowledge, which usually includes the contact process FAQ, customer high-frequency questions, and agent gold medals accumulated in the early stage of the enterprise. Words, etc. However, these data are usually lagging behind and in urgent need of aggregation and tailoring. They can no longer help agents give customers the latest and most accurate answers, nor can they support enterprises to complete effective operational analysis and optimization.
The second category is corporate documents. In the process of development, enterprises have accumulated a large amount of product manuals, equipment maintenance guides, supplier management systems and other contents. This information is increasing and iterating year by year. In the past, it could only rely on manual memory and recording of this massive and in-depth information, which made it impossible to Effectively respond to customer problems in actual business. The third category is session information. A large amount of session information will be generated in scenarios such as corporate pre-sales consultation and reception, sales opportunity conversion, and after-sales service. This information often hides huge commercial value. These three types of corporate knowledge connect corporate marketing, sales, service and other business scenarios, and are also important areas for the implementation of innovative technologies. However, most companies have not found effective solutions in this area. The emergence and application of large models provide new ideas for enterprises to solve such problems.
As we all know, the implementation principles of large models have already incorporated technologies such as machine learning. As a result, the relevant knowledge accumulated by enterprises during business implementation can be directly used to train AIGC capabilities that fit their business scenarios. , which can more flexibly cope with various scenarios in business implementation.
Before the emergence of large models, related solutions based on "knowledge base" were usually used in such scenarios. For example, the previous conventional customer service robots can only solve the relevant problems involved in the content of their "knowledge base". For communication problems that are not included or are slightly different from the included content, they can often only respond mechanically and answer questions incorrectly. Users in the scenario create a poor communication experience and are unable to solve actual problems. As Tian Runyun said, "The emergence of large models has provided a reliable 'connector' for corporate knowledge and customer contact, and has found a new breakthrough for corporate growth."
With a single revenue model and many industry challengers, what is the future for Tianrun Cloud?
According to a data from the Chishi Consulting Report, Tianrun Cloud is China's largest public cloud customer contact solution provider in terms of revenue in 2020, with a market share of 10.2%. Although the market in this field continues to be optimistic, as competition in the industry gradually intensifies, there may be variables in Tianrun Cloud's long-term performance.
On the one hand, of Tianrun Cloud’s overall revenue, revenue from the “intelligent contact center solutions” business accounts for as much as 81%, and the business structure is relatively simple. On the other hand, traditional giants from the Internet and related fields are also continuing to increase their investment in this business field.
Among them, at the 2023 Tencent Global Digital Ecosystem Conference on September 7, Tencent Qidian announced a number of latest developments and mentioned that it will "improve the efficiency and user experience of Tencent Qidian intelligent customer service management based on AI large model capabilities." In June this year, Tencent Qidian released a new generation of intelligent customer service around customer service scenarios, and by introducing large model capabilities, it improved the three levels of dialogue interaction, manual assistance and knowledge construction.
iFlytek, which has profound experience in the field of intelligent voice and artificial intelligence, also released a new generation of intelligent customer service products integrating iFlytek's Spark cognitive large model - iFlytek's Spark Intelligent Customer Service on June 30 this year. In addition to challenges from industry giants, cloud communication service provider and New York Stock Exchange-listed company Ronglian Cloud also released the "Red Rabbit Large Model" and "Generative Integrated Intelligent Customer Service Platform" products on July 8.
What is certain is that in the future Tianrun Cloud will face a market that is full of competition but is still in the growth stage. In the face of users’ increasing “intelligent” needs, the performance of large models and the ability to implement them in business scenarios will become the core capability that affects its future market performance. At the same time, Tianrun Cloud should also expand its business horizontally with the "intelligent customer contact center" business as the core. After all, facing the "one-stop" service solutions of enterprises such as Tencent Qidian, Tianrun Cloud's business segment is currently unable to achieve the same goal. Point to surface.
At present, the "Battle of Hundreds of Models" has not yet been settled. Tencent officially released its self-developed "Hunyuan Large Model" on September 8 and officially opened it to domestic enterprises through Tencent Cloud's API. From "should enterprises have their own large models" to "the last mile of commercialization of large models", the industry's understanding of large models is gradually becoming mature.
Since there is no killer application for C-end users in China, the topic of “large models” has become less popular recently. However, within the industry, the exploration and discussion of commercialization of large models has always maintained a high level of enthusiasm.
What value can the application of large models bring to enterprises? Especially in the field of enterprise services, which scenarios best match the characteristics of large models and have relatively stable value performance? It can be seen from Tianrun Cloud's semi-annual report that through the combination of large models and enterprise knowledge engineering, it will be a relatively typical application direction to integrate AIGC capabilities into relevant scenarios of customer contact solutions.
It can be seen from Tianrun Cloud’s business scenarios that enterprise users’ demand for more intelligent and efficient customer contact solutions is growing. Applying intelligence to enterprise service-related fields can help corporate customers reduce costs and increase efficiency, which meets the needs of enterprises in the current economic environment. Before killer applications appear in the C-side field, the commercial application exploration of large models in the enterprise service field will undoubtedly go relatively smoothly.
★ Follow the Data Yuan official account and reply "Tianrunyun 2023 Semi-Annual Report" in the background to download the full version of Tianrunyun's 2023 Semi-Annual Report.
Text: Nick / Data Monkey
