Shulou(Shulou.com)11/24 Report--

In the summer of 1956, computer scientist John McCarthy first put forward the concept of "artificial intelligence". Since then, AI has become a frequently mentioned topic in the scientific and technological circles.

In recent years, when AlphaGo beat go world champion Lee se-dol with a score of 4:1, and the born ChatGPT was adept at answering tricky puzzles, people were amazed that AI had developed to such a level, so the heat wave of AI once again swept the world and become the hottest science and technology hot spot.

Recently, at the "Zhiqi AI AIGC Industry Enablement Conference" sponsored by NetEase, a number of celebrities with deep understanding in the field of AI discussed the trend of AI enabling industry, the business opportunities generated by AI, and the strategy of AI-driven enterprise transformation. NetEase shared the practical experience of AIGC in improving organizational effectiveness, as well as the latest progress and achievements of AI landing in various business lines of smart enterprises, and released a large industry model "Shang he" focusing on customer service.

Let organizations evolve themselves like AI

Just as the emergence of fire broke the night, the emergence of electricity brought about a productivity revolution, and no one would doubt the significance of AI in this era. But major technological breakthroughs and applications have never been accomplished in a "miraculous" way, and many companies want to let AI cover all aspects of their business from day one after deciding to keep up with the heat wave of the AI era. The difference is that NetEase has adopted a "small-step, fast-running" pace, gradually integrating AI tools into the workflow, actively exploring and embracing the efficiency improvements brought about by new technologies, so that organizations can evolve like AI.

Ruan Liang, vice president of NetEase and general manager of NetEase Intelligence Enterprise

Ruan Liang, vice president of NetEase and general manager of NetEase, cited the first pilot project of the smart company as an example: based on customer Voice, optimize the GTM (Go To Market) trip of the product with the help of mature AI tools on the market. The standard process of new product commercialization will go through three stages: market demand research, product value verification (MVP) and product market fit verification (PMF), in which three key links are involved: external information acquisition, internal decision execution, and external market access.

In the external information acquisition link, based on the strong analysis, summary and induction ability of AI, we can purify the massive data and information from the front line, carry out targeted product improvement on the basis of these purified "customer Voice", and hit the customer pain point directly.

In the internal decision-making and execution link, AI enables business personnel to have professional ability, and business personnel directly participate in material output through AI, which, on the one hand, shortens the collaborative link, more importantly, allows the business perspective to better guide material output, and allows professionals to accomplish more professional things.

AI is also good at product marketing and delivery services. Through the rapid analysis of customer information to find the "right crowd", and batch generation of high-quality marketing copy and so on.

Finally, it runs through the whole process on this product, reducing the labor cost of the whole process from 852 to 675, and increasing the efficiency of each link by about 25% on average. More importantly, the greater significance of this 25% efficiency increase lies in reducing opportunity costs. If the improvement in efficiency enables the organization to "hit the south wall" more quickly, it also means that the organization can extract resources more quickly and invest them in more valuable places.

Scene first, be the polygon warrior of AI landing.

In the past few decades, every wave of AI has catalyzed the birth of great companies, so after the AI heat wave set off by ChatGPT, enterprises are no less enthusiastic about making a gorgeous turn with AI. After flocking to it and moving forward one after another, leaving a place of chicken feathers or a bunch of stars is often related to the attitude and choice of enterprises.

NetEase adheres to the mission of "gathering people on the Internet to create a better life through scientific and technological innovation." in fact, as early as before ChatGPT set off the AI heat wave, NetEase had already laid out AI for many years, and it was only after accumulated strength that he would frequently drop his son this year, such as the large model "Zi Yue" for the education field, the self-developed Chinese text pre-training model "Yuyan", the picture-and-text generation model "picture and text", and so on.

Following NetEase's pragmatic style of always putting the scene first and paying attention to landing, in the field of ToB, NetEase intelligent enterprises also hope to use the best technology and services to help customers grow endogenously. After AI serves the internal scene of the organization and realizes the self-evolution of the organization, NetEase Intelligent Enterprise will fully open its mature practice to customers through technology and services.

As we all know, game is one of NetEase's important labels. In fact, AI is already ubiquitous in many games, and some AI are guides for novice players, while others fight side by side with players. As the game becomes more complex, the competitive level of AI characters needs to be on a par with that of real-life players, but the AI is still too clumsy. Based on rich game training data and in-depth experience in the development and application of AI, NetEase Intelligence Company has developed a game AI agent, which can compete with top human players without losing ground, and supports high customization. it can train different intensity according to the data of different levels of players and adapt to different levels of confrontation. At the same time, the injection of AIGC ability makes the game AI agent have strong reading comprehension, reasoning and deduction ability, can generate content independently, understand the player's intention, interact with the player, cooperate with the player, and establish a deep social relationship, which gives the soul to the game AI.

In addition to games, I believe you must have had such an experience: in online meetings or other scenes with audio and video calls, live broadcasts and other media, noises such as keyboards, vehicle sirens and outsiders are constantly heard, masking each other's voice and affecting the quality and experience of the call. The AI noise reduction algorithm developed by Yi Yunxin, an expert network of converged communication cloud services owned by NetEase, can automatically identify and filter ambient sound components, and directional noise reduction of background voice, keyboard voice and other unsteady noise in noisy environment, the algorithm has reached the industry-leading level.

Whether it is the above-mentioned games or live streaming, or browsing social platforms and watching various video sites, there is still a selfless "AI judge" standing behind. Through AI and other professional technical means and business rules, NetEase Dun, a digital content risk control expert under NetEase Intelligence Enterprise, provides comprehensive digital content risk control products in the three major directions of content security, application security and business security.

The large model of "Shanghe" helps enterprises to build service flywheel

In this Wuzhen AIGC industry empowerment conference, NetEase intelligent enterprise released the "Shanghe" big model. Shanghe River comes from Laozi's "running water is not first, but the argument is eloquent." In the long river of commerce, when it flows like a stream, do not fight for the priority of the moment, keep your feet on the ground, step by step, and save your strength bit by bit.

Ruan Liang, vice president of NetEase and general manager of NetEase Intelligence Enterprise, said that in the thinking of the big model, NetEase intelligent enterprise insists on being pragmatic and locates the industry model in the vertical field based on its own advantages. As a result, Shanghe has become a "specialist" in the big model, focusing on the vertical field of customer service. In addition to using general data for pedestal model training, supervised fine-tuning (SFT) and human feedback-based reinforcement learning (RLHF) are also carried out based on the nearly 10 billion customer service corpus accumulated by NetEase Cloud Merchants.

Compared with the general model, although the requirements for cost input have decreased, Shanghe has a richer and more comprehensive "knowledge reserve" in the vertical field of customer service. Thanks to this, NetEase Cloud Merchants' new generation of intelligent customer service, which is equipped with Shanghe model, can handle more complex information, understand the context more accurately, have stronger abilities of logical reasoning, emotional analysis, information extraction, ambiguity resolution, and multilingual translation, and aim at the "rigid demand" of enterprises for intelligent customer service to reduce costs, increase efficiency and even increase income, and build a rapidly rotating service flywheel for enterprises.

End

"I can imagine a world of artificial intelligence in which we are more productive, we live longer, and we have cleaner energy." Computer science experts from Stanford University had such a yearning for AI.

With the passage of time, AI eventually attracted more and more believers, but also with its unique charm to dispel all kinds of skeptics, people began to invest more energy to describe a near future.

But on the way to the future, only with keen market insight, profound scene disassembly ability and high-quality technology fusion ability can we build their own moat, NetEase intelligent enterprises will continue to make efforts to this end.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.