After the "Battle of the Models," discussions surrounding the implementation of AI applications in 2024 have intensified, particularly in the e-commerce sector. This area appears to be the fastest adopter of generative AI, benefiting from its intrinsic advantage and proximity to revenue. E-commerce offers significant opportunities for AI to transform various aspects of the industry, including manufacturing, operations, and marketing.
"The era of AI in e-commerce is just beginning, presenting both opportunities and challenges for everyone," said Jack Ma at the end of last year, as Alibaba outlined its user-centric, AI-driven growth strategy. Major players such as JD, Pinduoduo, and TikTok have also developed their own AI e-commerce strategies, while international giants like Amazon and Google are advancing AI-related projects.
Recent updates from Alibaba International show a rapid pace in AI deployment. To date, AI has been utilized across over 40 different scenarios, benefiting approximately 500,000 small and medium-sized businesses and optimizing around 100 million products. The demand for AI services among merchants continues to rise, with request volumes doubling every two months, and recent daily averages surpassing 50 million. "Our primary focus in the past six months has been on rapidly applying AI," stated Zhang Kaifu, Vice President of Alibaba International Digital Commerce Group and head of its AI initiatives. The AI Business team, established in April of last year, has grown to over 100 members and focuses on practical AI applications.
Alibaba International Digital Commerce (AIDC) is one of Alibaba's fastest-growing divisions, boasting more than 300 million active annual consumers. The division encompasses international retail businesses such as AliExpress, Lazada, Trendyol, and Daraz, as well as wholesale operations in international markets. Alibaba employs a multi-platform strategy, requiring a shared AI infrastructure while maintaining some independence in data centers and model training. This setup provides ample space for the AI Business team to innovate.
Zhang noted that AI's greatest value in cross-border e-commerce lies not in replacing human roles but in enhancing metrics such as click rates and conversion rates, ultimately reducing costs and increasing efficiency. As the demand for overseas expansion surges, managing costs has become increasingly critical. Selling products internationally involves navigating complex processes such as creating product titles and descriptions, photographing items, generating advertising materials, and managing customer service and returns.
In these scenarios, AI excels in customer service and marketing, providing double-digit improvements in effectiveness. Additionally, it can boost conversion rates, click rates, and consumer satisfaction by 1% to 30%, depending on the depth of application in distinct contexts. However, developing AI systems requires substantial investment, and even a giant like Alibaba must strike a balance among performance, efficiency, and cost constraints. Determining priorities within various scenarios and assessing their potential business impact is crucial.
Currently, most of Alibaba International's AI efforts revolve around optimizing workflows. When it comes to product content, categorization primarily occurs by country or region, making it challenging to achieve personalized experiences for individual users. Nevertheless, there is significant potential for AI in the future—both in enhancing supply chains and innovating product forms.
An emerging startup, Daydream, is exploring innovative AI-driven shopping experiences. Recently receiving $50 million in seed funding, Daydream aims to revolutionize traditional e-commerce search and recommendation by using AI to combine natural language and images for more personalized shopping results. For example, users could search naturally for apparel by saying, "I'm attending a wedding in Paris this fall. Can you recommend some outfits?" They can even upload reference images for tailored suggestions. Although the platform is not yet fully operational, it is set to launch a test version this fall.
Reports indicate that Daydream currently hosts over 2,000 brands, focusing primarily on niche markets such as Altuzarra, Jimmy Choo, and Alo Yoga. While Daydream sets a new standard in AI-driven e-commerce searches, it faces challenges—its recommendations are limited to the brands it features, leaving countless fashion items undiscovered by AI.
In China, the models for search, recommendations, and advertising have become foundational to the internet platform landscape. Whether large models can fundamentally transform this approach poses both significant costs and a shift from long-standing business methods. Zhang acknowledged that transformation in the industry is gradual, as established methods are deeply rooted. However, he emphasized that developing new product expression techniques at the algorithmic level is a compelling topic.
At present, Alibaba International’s AI focuses on enhancing efficiency within existing business models and workflows. With the scaling of AI in cross-border e-commerce, technology is expected to increasingly influence business models, product strategies, and operational processes.
Ranking
