By akademiotoelektronik, 06/10/2022

Artificial intelligence, Alibaba's lethal weapon to supplant Amazon

According to Grégory Pouy, marketing consultant quoted by Petit Web, Amazon is “lag behind” the Chinese. And to quote an example: “When you buy a brush on Alibaba, the site recommends painting lessons. Whereas on Amazon, once you have purchased a brush, the site recommends others.”

It is true that the recommendations that can be found, for example, on AliExpress, a site specializing in wholesale, are quite impressive: often, what is offered to you matches what you would probably have looked for on your own. Personally, I haven't found in my experiences with Alibaba that these recommendations were so different from Amazon's - that I find equally fair and close to my needs.

But who currently offers the best suggestions is not the most important (by the way, if you have an opinion on the question, do not hesitate to mention it in the comments). The idea, implicitly, is above all that AI could be a way for Amazon's competitors in e-commerce to gain the upper hand. So, would it be possible? Could Alibaba one day defeat Amazon with its artificial intelligence?

Amazon, a war machine

In 20 years, Amazon has become the benchmark for e-commerce sites in France, Europe and the United States. When we talk about “web giants”, we rarely think of Amazon, focused as we are on Google, Apple and Facebook, but the firm of Jeff Bezos could well one day overtake its 3 colleagues.

According to ScottGalloway, professor of marketing at the NY University School of Business, Amazon will stand out for several reasons. First, because it is the dominant online store, representing 45% of online commerce in the United States, to the point that the postmen sometimes have the impression of working for the firm. Amazon even dominates online searches: thus, 55% of searches (the figure increases each year) on a product to buy start directly on its site, and not on Google.

Secondly, because its enormous market capitalization (700 billion dollars, soon to be 1000) allows it to invest massively in all strategic sectors: video on demand and Netflix-style fiction production with Amazon Prime Video, voice assistants with Echo (which stands up to GoogleHome and Apple's HomePod), physical commerce and food with the chain of organic supermarkets Whole Foods (acquired in June 2017) and its Prime Now service, or cloud computing services for businesses and individuals with AWS (Amazon Web Services). A very lucrative little “empire”.

Alibaba, the Chinese outsider

Faced with this war machine, only one real competitor: the Chinese Alibaba, which is beginning to expand in the world and also invest everywhere, with a equally large market capitalization ($500 billion). Its e-commerce sites for individuals, AliExpress and Taobao, are increasingly popular in Europe, particularly because what they offer for sale are very inexpensive products. But as it stands, Alibaba is only a "marketplace": it connects sellers and buyers, but does not sell its own products, does not manage inventory, and does not ensure deliveries - unlike Amazon which delivers “in one click” and in 24 hours with its premium service, Prime.

Alibaba is trying to track Amazon as much as it can by diversifying and globalizing. In addition to AliExpress in Europe and Taobao, the BATX (equivalent to GAFA in Asia) has e-commerce sites for Chinese and international brands (Tmall and Tmall Global), service sites such as Alibaba Cloud, Alitrip or Flizy for travel, and Alipay, a remote payment service. For a long time, he only invested in Southeast Asia, but since 2014, he has also invested in India, where Amazon is present. Alibaba is also apparently starting to eye delivery, investing in SingPost, Singapore's postal system, and acquiring Cainiao, a Chinese logistics network. Last summer, it also launched a competitor to the Amazon Echo and Alexa: the Tmall Genie X1, which has its own voice assistant, AllGenie, and which helps its user in the purchasing process.

AI, a competitive advantage

But despite these massive investments, and the fact that Alibaba overtakes Amazon in number of customers (454 million against 310, in 2017), the Chinese group is still struggling to break through in the West, where it is still relatively unknown, in addition to delivering more slowly than the American company. Hence the interest of artificial intelligence, as a competitive advantage.

Personalization is the key to e-commerce. By subtly suggesting products for the user to buy based on what they have already purchased, or their visits to the site, the AI ​​(whether Amazon's or Alibaba), gives it a very useful “boost”. And when the buyer appreciates a useful recommendation, he remembers it and generally tends to return to the e-commerce site in question.

Thanks to its product recommendations, Amazon recorded a 34% increase in sales in the last half of 2017. Easy to understand, therefore, why the GAFA has been relying on AI for some time now: for example, it is currently opening research centers in Barcelona and Germany, in order to develop more and more machine learning and processing algorithms. natural language (NLP), which, coupled with its huge database of consumer purchases, will allow it to more accurately predict the items its customers might want to buy. AI also allows Amazon to operate its personal assistant ,Alexa, as well as to optimize its delivery activities. Amazon's patent on one-click payments expired last fall - and it's no doubt to keep pushing its competitors that the firm recently filed "several AI and Machine Learning patents," according to Business Insider.

Alibaba, deep into AI

Alibaba seems ready to take up the challenge: it plans to invest 15 billion dollars in R&D, mainly in the field of AI, between 2018 and 2021 ( with the creation of 7 research labs, all over the world). And he would be wrong to deprive himself of it, because apparently Amazon's AI is still far from perfect. Researchers at the University of Toronto tested its product recommendation algorithms: “AI predicts exactly what we want to buy about 5% of the time. In other words, we buy about one out of every 20 items he recommends,” they state. Suffice to say that despite the impression of ordinary users of having fine suggestions in front of them, Amazon still has some work to do.

The Chinese e-commerce giant uses, like Amazon, a massive customer database (which comes from its many sites, but also from Alipay), coupled with artificial intelligence. Its Alibaba recommendation system, “E-Commerce Brain,” uses real-time online data, along with artificial neural networks and algorithms, to build models that allow it to predict what consumers want. AI, these predictive models are updated for each Internet user, based on their past purchases, but also on their browsing, their comments, their “favorites” and their preferences (product categories, brands, etc.), as well as their behavior shopping millions of other users of the site. According to Alibaba, the use of this technology has already enabled it to increase its conversion rate by 20% last year.

Alibaba places its pawns

Just like Amazon with Web Services, the Chinese group also launched its own AI platform, DT PAI, this year, intended for developers and companies that use its online sales sites , in order to help them “predict user behavior”. By predicting purchasing trends, DT PAI should enable sellers to improve their products, as well as their inventory and delivery management.

According to Bryan Ciambella of consulting firm B Capital, 75% of US consumers “believe that no one can provide a better online experience than Amazon”. Alibaba is forbidden to invest in the United States for the moment, but for how long? The Chinese are placing their pawns, and are preparing to flood Europe first. In the latest news, in France anyway, the Asian e-commerce group is starting to correct one of its weak points, delivery times. According to Le Monde, since July 2017 it has been testing delivery in France from a warehouse located in the Paris region. Alibaba would also be looking for storage space to rent, and would discuss with “logistics partners”. On the AliExpress site, 24-hour deliveries are also offered.

It should be noted that Chinese progress in the field of AI is quite impressive: in January 2018, one of its research centers, the Institute of Data Science of Technologies (iDST), developed a model artificial neural network that outperformed humans on a reading and comprehension test designed by Stanford University. This breakthrough in the field of NLP most likely gave Amazon cold sweats. The task is difficult, but nothing is impossible, and if it is well designed (and actually recommends painting lessons), Alibaba's recommendation AI could very well make the difference.