Since its listing on the main board of the Hong Kong Stock Exchange on June 30, 2022, Tianrun Rong has continuously increased its investment in R&D in AI, and established the product development goal of reconstructing customer contact systems and processes with "AI native" thinking. In the past year, Tianrun Rongtong has launched a number of innovative AI products. These innovative products have driven rapid growth in the number of customers and the scale of revenue.
Since the beginning of this year, the rapid development of large language models represented by ChatGPT has injected strong impetus into the landing of artificial intelligence in various application scenarios. At the same time, the big language model also activates the imagination of customers, prompting customers to actively participate in the innovation of AI applications, which is an important innovation force that will accelerate the "qualitative change moment" of customer contact intelligence.
Wu Qiang, founder and CEO of Tianrun Rongtong
PART 01.
From self-service to intelligence: the AI evolution of customer contact
With the rapid development of AI technology, traditional "self-service" customer contact is undergoing a revolution. In the past, customers had to face template-based self-service systems to interact with businesses. This makes it difficult to meet customer needs, and there is always a gap between enterprises and customers. However, innovations such as the Big Language Model in AI technology have revolutionized customer contact, identifying a customer's intent, emotion, and needs and giving a personalized, targeted response.
In the view of Tianrun Rongtong, the development process of customer contact from self-service to intelligence can be divided into four stages:
1.0 era: basic self-service
Jump the service by pressing the button, select the required service according to the prompt sound, and if the configuration is not overridden, you can only transfer to the human service.
2.0 era: based on rich keyword matching technology
At this stage, it can be understood that when the keyword preset by the voice or text trigger system, the business jump is richer than that supported by the button method, but the maintenance cost is higher, and when the keyword changes, it cannot be successfully matched.
3.0 era: semantic understanding based on deep learning
Through deep learning technologies, such as convolutional neural networks, recurrent neural networks, etc., a large amount of data is analyzed, so that AI customer service can accurately understand the user's intention, so that the customer can feel that it has a sense of realism close to human intelligence. In addition, at this stage, technical means such as multimodal learning can be applied to realize natural language understanding, dialogue management and other capabilities, and provide users with multi-dimensional experience services.
4.0 era: large language model vertical industry application
The scale of parameters is larger (hundreds of billions or trillions), which can use industry data and knowledge to provide more accurate and efficient solutions, better meet the needs and expectations of users in a certain field or scenario, better understand the special needs of users, and provide more perfect AI solutions.
PART 02.
From "quantitative change" to "qualitative change"
The continuous upgrading of AI technology is a process of "quantitative change", and the early models are small in scale and low in complexity, and can only handle some simple, specialized, and goal-specific tasks, which have great limitations. With the continuous improvement of computing power, the larger scale of AI models, and the more complex structure, AI technology has the ability to continue learning, in-depth analysis, and generative, and it is this kind of technology "quantitative change" that directly promotes the "qualitative change" at the application level and opens the "qualitative change moment" in the field of customer contact, which we define as the following characteristics:
- Works right out of the box
Early stage: The application of an AI model needs to go through the process of data preparation, model evaluation, and model tuning, which usually takes a long time, has high comprehensive cost, and is slow to present value.
Qualitative change moment: Based on massive parameters and content reserves, the cold start capability of enterprise intelligent services can be improved, and after deployment, it can be put into business scenarios, with low comprehensive cost, fast online launch, and rapid presentation of business value.
- Multi-modulus
Early stage: Different business scenarios require the deployment of independent AI models, such as multilingual translation, text-to-speech conversion, quality inspection, and robotics, which lead to high integration costs, high maintenance complexity, and siloed effects, resulting in fragmented information, which cannot maximize business value.
Qualitative change: A set of large-language model technology can realize a variety of task processing, low deployment cost, data sharing and comprehensive decision-making, which can maximize business value.
- Emotional interaction
Early stage: The need to manually formulate rules and policies, unable to personalize the needs of different customers, cold and blunt robot service capabilities, unable to fully release the value output of artificial agents.
Qualitative change moment: Through the real-time analysis of the conversation process, it can effectively monitor the customer's emotional state, communication intention, etc., and make emotional superposition such as appeasement, guidance, and approval in response speech, maintain emotional resonance with customers, improve semantic and emotional understanding ability, improve the accuracy of consultation response, customer satisfaction, and reduce customer complaint rate.
- Autonomous evolution
Early stage: Through manual extraction, creation, manual sorting and screening of various documents, pictures, videos and other information, the comprehensive maintenance cost is large, the update speed is slow, and the knowledge applied to the front line is seriously lagging behind, which cannot effectively support the service and lead/opportunity conversion process.
Qualitative change moment: the integration of all kinds of knowledge documents/pictures/videos and other knowledge base information can be automatically extracted, analyzed, aggregated, reasoned, retrieved, generated, etc., the maintenance cost is extremely low and automatically completed by the machine, and the knowledge base is always up-to-date, which can provide strong support for the front line and improve customer satisfaction and business goal achievement rate.
Based on the in-depth insight into the big language model, Tianrun Rong has upgraded itself, completed the comprehensive integration of AI products and large language model technology, and opened the "qualitative change" moment of Tianrun Rong.
PART 03.
With AI native as the engine, what kind of new experience will big language models bring to enterprises?
In the past few months, from ChatGPT to domestic Wenxin Yiyan, Tongyi Qianwen, Xinghuo cognitive big model, 360 wisdom brain, etc., major manufacturers have launched their own general large language model products, Tianrun Rongtong is more concerned about how to quickly realize the real landing of large language models in customer contact scenarios, bringing more value to enterprises.
Wu Qiang said that Tianrun Rongtong will be driven by AI technology to build a new customer contact experience of "human-machine integration" to help customers achieve the goals of efficient assistance, close collaboration and deep insight.
Tianrun integrates human-machine integration capability matrix
Through the integration of self-developed AI products and large language model technology, the customer contact field will bring three major value enhancements to enterprises:
- Efficient assistance to create a new customer experience that integrates man and machine
In the past, when providing customer service, enterprises were received through a large number of labor, and the emergence of robots can effectively free labor from low-value work. Let humans focus more on solving complex problems.
While the capabilities of robots have advanced significantly in many areas, "human assistance" and "assisted decision-making" are also crucial in business scenarios. In the past, the training cycle of professional customer service personnel was long, the cost was high, and the low success rate was also a long-term problem faced by enterprises. With agent assistants, the professionalism of newcomers can be greatly improved. For example: according to business rules for telephone navigation, customer consultation content automatically matched with the best speech recommendation, automatic filling in work orders, etc., so that the work efficiency and professionalism of manual customer service can be significantly changed, so as to enhance the professional image and customer experience of the enterprise.
In the process of customer service, enterprises accumulate a lot of valuable session information. However, this information is often not used effectively. Through in-depth conversation analysis technology, we can effectively extract and analyze the key information generated by customers in the dialogue process, and help enterprises improve service quality and product promotion strategies.
- Close collaboration to build a knowledge foundation for the rapid development of enterprises
In the three core business scenarios of marketing, sales and service, a large number of knowledge bases are inseparable as an important guarantee for business development. Previously, the knowledge held by enterprises was often scattered among different departments and different people, making it difficult to form a knowledge base that effectively supported business development. Traditional knowledge management requires a lot of time and labor costs, and it is difficult to respond quickly to customer needs. However, through the vertical application of large language models, enterprises can easily achieve efficient knowledge management, including one-click FAQ expansion, document knowledge extraction, document-based learning and answering, and automatic knowledge learning in the contact process, so that robots and artificial agents can maintain the accuracy and real-time nature of knowledge in the customer contact process, and transform knowledge results into performance value.
Traditional multi-department business collaboration scenarios often have the following shortcomings: the process is cumbersome and inefficient, and customer service personnel need to manually create and process work orders, resulting in slow processing speed and long feedback cycle, which seriously affects customer service efficiency. Through the context understanding ability of AI conversation analysis, combined with automatic work order clustering and content analysis technology, it can timely capture customer needs and automatically generate work orders, automatically assign processing personnel, automatically update progress, and notify customers in real time, thereby optimizing traditional multi-department business collaboration scenarios, improving the efficiency of work order processing and fully closed-loop service efficiency, and improving the quality and effect of customer service.
- Insights to turn ever-changing markets into predictable opportunities
At present, the common problems encountered by many enterprises in the process of operation are largely due to the problems of difficult data connection, large volume and difficult analysis, long output of multi-department and multi-role business analysis reports, and difficulty in mining business value, which have caused a great waste of data resources. Relying on the integration of AI products and large language model technology, data can be further aggregated, cleaned, mined and visualized, such as displaying customer voice, line intention, high-frequency questions, gold medal skills, service level, service efficiency, customer satisfaction, customer sentiment, etc. The effective awakening of a large amount of data can provide an optimization basis for the business strategy of enterprises and help them enhance their market competitiveness.
PART 04.
It is the beginning, but also the future
AI technology represented by large language models is not only a competition with technology for service providers in various fields, but also a race against time. This not only tests the service provider's understanding of new scenarios and new needs, but also tests who can create more unimaginable surprises for customers through rapid iteration of technology application, data accumulation and service capabilities.
We have reason to believe that the big language model is just the beginning, and the integration of AI and industry will have a better future.
"WeChat search for "Tianrun Rongtong", send "big model" in the background, and apply for a free trial at the end of the article!"