E commerce changes , AI restoration

Visual China.

Brain polar body.

There are many e-commerce practitioners who still stick to their posts during the Chinese New Year. This year, Jingdong and Xiaohongshu, as the only two major factories in the 2024 CCTV Spring Festival Gala, have a sense of presence in front of the national audience. Among them, JD.com is a familiar traditional e-commerce giant, and strives to activate new growth through the Spring Festival Gala red envelope activity. Xiaohongshu simultaneously launched the "same style of the Spring Festival Gala" in the live broadcast of "Everyone's Spring Festival Gala", making great strides into the field of e-commerce.

This situation of "old and new scuffle" has been the "main theme" of the e-commerce industry for more than a year, and will become more and more intense in 2024.

In fact, as early as January before the holiday, when various platforms were in full swing, a big discussion broke out about "new and old e-commerce".

Some consumers, Internet product managers, industry analysts, etc., believe that traditional e-commerce companies such as Ali (**Tmall, Yitao, etc.) and JD.com have serious "path dependence", continuing the usual marketing methods in the past - "New Year's Festival". Only Pinduoduo shows "normal delivery", which fiercely grasps the pain points of users who are closed for express delivery during the Spring Festival.

The opposite side believes that this wave of Pinduoduo is "overreputation" and "embarrassing", but there are differences in the promotional copy, Jingdong and ** also have "Spring Festival does not close" activities, and no one will completely lie down.

The fierce quarrel shows that multiple e-commerce platforms are "rolled up", no one wants to lie down to make money, and consumers are happy to see it. At the same time, the public has also complained about the lack of innovation of traditional e-commerce for a long time, "hating iron does not make steel".

This can't help but remind me of the end of last year, when Pinduoduo's market value surpassed Alibaba's, Ma Yun made a rare statement on the intranet, "The era of AI e-commerce has just begun, and it is both an opportunity and a challenge for everyone."

The words are still in my ears, but the problem is that AI may be an unfamiliar concept to many people, but it is not a new thing for "digital native" e-commerce companies, as early as 2013, ** proposed the algorithm of "thousands of people and thousands of faces". In Pinduoduo's prospectus a few years ago, it was mentioned that "Pinduoduo is driven by a distributed intelligent ** network (distributed AI)".

After the reform and reform of the law, Li Hongzhang said: "If the old law can be rich and strong, then China will be strong for a long time, why wait for today?" "With the same logic, if AI can stay ahead, then traditional e-commerce has not been weak in technology, so why will it be caught up by rising stars?

From "e-commerce + AI" to "AI e-commerce", in the cruel stock era, what the e-commerce industry needs is not only a technical upgrade from machine learning to large models, but a more thorough "change".

At present, leading e-commerce platforms and AI technology companies have gathered heavy troops in the field of "AI e-commerce".

"Shoucheng" track:Ali and Jingdong must keep their head positions in the field of e-commerce. Taotian Group has just established a complete AI team, bringing together more than 20 teams originally responsible for AI business into 4 teams, which are responsible for Alimama, C-end consumers, B-end merchants, and industry-specific applications. Based on the Yanxi industry model, JD has launched JD Cloud AIGC content marketing platform, Yanxi digital human platform, and a series of AI tools for JD Retail and external customers.

"Enterprising" track:Pinduoduo, ByteDance, Kuaishou, Xiaohongshu, etc., have their own e-commerce businesses, and they have also launched AI tools one after another.

"Crossover" track:, Tencent and other basic model manufacturers have also squeezed into the e-commerce battle with AI and launched targeted technical solutions.

In general, unlike the previous "e-commerce + AI" that applied machine learning and small models to business, this round of AI boom set off by large models has an unprecedented impact on the e-commerce industry, and the competition is extremely fierce.

The first reason for this is the external crisis. The overall growth rate of the e-commerce industry has slowed down, entering the stage of competition in the stock market, the contradiction between traditional e-commerce and new platforms has also deepened step by step, and at the same time, consumers' behavior and decision-making habits have also changed greatly.

The second is the internal foundation. E-commerce enterprises have been built on the Internet and mobile Internet from the beginning, and are one of the earliest and most mature industries in digitalization, with a large number of businesses being driven by data, which provides a good foundation for the introduction of AI.

Driven by internal and external factors, the e-commerce industry has ushered in a vigorous AI revolution.

The powerful comprehension and generalization capabilities of large models make them more subversive than previous machine learning in creating, generating, and decision-making tasks, and can transform every link of the e-commerce industry chain.

Let's take a look at the novelty that AI brings to e-commerce through the four core steps of transactions:

1.Build a store to operate.

The premise of online business is to have a display platform for goods and a channel for carrying traffic, and "online store" is a necessary entrance. For a business, daily business activities such as opening a store, shooting, art, copywriting, and launching new products have a long cycle, many trivial matters, high costs, and a large number of repetitive operations, which make employees miserable. The generation capabilities of large models and AIGC tools can help merchants open stores quickly, automatically generate product materials**, save costs such as modeling, shooting, and post-production, write detailed page copywriting, optimize titles, etc.

The threshold for online stores has been greatly reduced, and AI can help young entrepreneurs, agriculture, forestry, animal husbandry, sideline fisheries, and other traditional industries with low operational capabilities to quickly enter the ranks of e-commerce.

2.Creative Marketing.

With a store, does the traffic come from? With the younger age of consumers, shopping habits have also undergone great changes, and there are two main growth methods with higher conversion rates: one is based on content marketing, and interest determines purchase; The second is consumption based on anchor interaction, and the decision-making link is shorter.

For merchants, content marketing involves product selection, copywriting production, short production, and content delivery, and every step is accurate, so as to create a "hot model" and achieve business explosion. Alimama has aggregated and upgraded multiple marketing tools such as through train, gravitational Rubik's Cube, and Wanxiangtai to "Wanxiangtai Unbounded Edition", intelligently generating advertising content according to target users, and then intelligently placing them. After the end of a marketing campaign, AI intelligent analysis is used to review and summarize to provide decision-making help for novice merchants.

The traffic of live streaming is fierce and the growth is strong, but a team of live anchors, including anchors, operations, central control and other professionals, is something that many small, medium and micro businesses cannot afford. Digital human live broadcast has become a better alternative, which can be broadcast live 24 hours a day, interact with users at any time, and do not miss the opportunity to make a deal. And only a virtual digital person and one operation are needed, and the cost is greatly reduced. At present, digital human is one of the capabilities that various e-commerce platforms and AI large model manufacturers are focusing on upgrading. According to Jingdong War Report, during the 618 period in 2023, the number of merchants launched by Yanxi Digital Human increased by more than 5 times compared with last year's Double 11. E-commerce's Huiboxing 7*24h plan for the sun never sets, and the GMV of a single live broadcast in some high-quality live broadcast rooms is as high as 3 million.

3.Interactive Services.

Attracting new users through content planting and digital live streaming also requires a series of services such as reception, consultation and shopping guide. At present, many e-commerce platforms have made provisions on the response time of customer service to improve the consumer experience, which also brings a lot of pressure to platform merchants and manual customer service. Based on the large model, AI customer service and AI shopping guide can understand users' questions and provide Q&A services. This is also the focus of AI e-commerce, including JD.com's AI intelligent customer service "Jing Xiaozhi", *'s "**Q&A", the preferred intelligent assistant "Xiaoyou" launched by e-commerce, and Xiaohongshu's AI intelligent note-taking assistant.

4.Logistics fulfillment.

Logistics fulfillment is a relatively heavy part of the e-commerce service chain and directly affects the user experience. Whether the receipt and delivery are fast or not, whether it is good or not, the user's experience will determine the brand reputation and repurchase rate.

For some e-commerce platforms with logistics systems, AI technology can be used to optimize the management of the first chain and improve the efficiency and quality of fulfillment, which can form a differentiated competitive advantage.

For example, the AI of the Yanxi model empowers JD.com's digital intelligence chain to improve the efficiency of commodity circulation; The official website of Cainiao also shows that a large number of Alibaba's AI technologies have been applied, including intelligent scheduling algorithms, intelligent address recognition, box cutting algorithms, etc., to improve the efficiency of logistics operations.

All in all, from the opening of new stores to the completion of contract fulfillment, AI has blossomed everywhere in the upstream and downstream industry chain of e-commerce, showing the vitality of comprehensive transformation.

All in all, AI e-commerce is full of imagination. But just like the growth of all new things in history, they have to go through a long period of maturity, facing the blockage from conservative forces, and AI e-commerce is no exception.

From the perspective of AI itself, the maturity of the technology still needs to be polished.

AI content generation for e-commerce needs to be improved in terms of industry segmentation, content compliance, and image generation effects. Blindly pursuing AI to reduce costs and increase efficiency, and the content is not effective and rushed online, it is easy to arouse consumer disgust, and the gains outweigh the losses.

At present, the "first batch of AI e-commerce victims" has appeared on social networks. Some businesses use AI-generated model images to be very different from the real thing, which does not truly reflect the product information, but increases the difficulty of consumers' decision-making, and hurts the user experience and trust.

Low-quality and distorted AI content may accelerate the loss of users and the decline of word of mouth, so e-commerce platforms must be cautious and formulate corresponding feedback and normative mechanisms to restrain them.

Intelligent customer service and shopping guides also have similar problems. Consumers need to make consumption decisions seriously and responsibly introduce and suggest, and the "illusion" problem of large models brings uncontrollable. Cao Peng, Chairman of the Technical Committee of JD Group and President of JD Cloud Business Division, previously said to **, "Users will laugh when Wenxin Yiyan and Xinghuo produce an wrong answer, but in serious business practice, a mistake means immeasurable losses".

Digital human anchors also have the problem of immature large-scale application of technology. It must be admitted that the current digital human anchors have limited interaction capabilities, and most of them can only carry out simple explanations and product discounts, and cannot communicate with users in depth, and it is difficult for consumers to establish an emotional connection with the digital human anchors, and they will be aesthetically tired after watching it a few times.

In the final analysis, AI is still a "fist and legs" for e-commerce, and it needs to be updated and iteratively optimized in technology, as well as the development of more AI-native applications.

Compared with the technology itself, the inertial force from the organization may be the biggest obstacle to the transformation of e-commerce, and it is also more difficult to touch.

Take the Spring Festival promotion as an example, AI may not be able to write "normal delivery" copywriting through data and competition analysis, but compared with the "New Year's Festival" has done this many times, the staff is proficient in operation, and the implementation effect is also good, a new plan with unknown effect, difficult performance, and high risk cost can be smoothly promoted internally? Can multiple departments work together? What about merchants and logistics systems?

The rudder of AI is ultimately in the hands of people. All reforms that touch the root must be promoted from top to bottom by the "number one project."

We have visited many retail enterprises with digital transformation, such as a top 500 domestic down jacket brand, which has set up a "digital special class", composed of persons in charge from production to marketing, to coordinate the interests and enthusiasm of all parties, so as to promote the full implementation of digital and intelligent technology internally.

The same is true for AI e-commerce, which not only needs to upgrade technology, but also to carry out comprehensive and in-depth transformation in many aspects such as organization and management, industry openness, and channel relationships.

For example, in order for AI tools to bring productivity growth to merchants, they must be fine-tuned according to industry characteristics and exclusive data, and do a good job of productization polishing, so that novice merchants can easily get started, and the quality and cost of AI-generated content are controllable. This kind of AI tools and development work will be very large and trivial, and it is impossible to do it all by relying on the e-commerce platform itself, which requires a group of AI developers.

So, do e-commerce platforms have an open mind to support the AI developer ecosystem and third-party service providers, and can they maintain the sense of boundary that does not touch user data and does not do user business?

If all AI tools and products are developed in-house, is there sufficient development force in-house, and is the cooperation between teams efficient? Is technology research and development closely integrated with the implementation to solve practical problems in business operation? After the launch of AI products, does it provide merchants with sufficient support such as platform environment, easy-to-use tool chain, and storage and computing network resources?

There is no doubt that e-commerce is the best training ground for AI technology, and AI is an indispensable weapon in the e-commerce industry. Combining the two, AI e-commerce will be one of the most anticipated innovations in 2024.

But in the end, which companies will be sent to the top of the tide of the times by new technologies depends on whether they are thoroughly believed in and forcefully executed.