今年618 迎来AI购物“大乱斗”

During this year's 618 mid-year shopping festival, all major Chinese e-commerce platforms rolled out AI-powered shopping features, and the industry has signaled that the AI e-commerce era will arrive by 2026. Below is key practical takeaways for general consumers:

1. Leading AI shopping tools each have distinct characteristics: Qianwen generates lively, conversational recommendations, organizes quick comparison tables for shoppers, and actively follows up to refine user requirements. Doubao offers objective, rational suggestions, explicitly notes product pros and cons, frames analysis from the user's perspective, and supports more informed decision-making. JD AI Shopping generates a larger volume of product recommendations, but suffers from muddled category logic, low recommendation accuracy, and lacks follow-up to clarify user needs.

2. Practical tips: To get more accurate recommendations, start by stating your general needs, then refine your requirements based on the AI's follow-up questions. Even when you ask non-shopping questions such as hiking guides, AI can proactively identify related shopping needs and provide one-stop reference. Consumers should note that current AI shopping tools still cannot compare prices across platforms in real time, have inconsistent information accuracy, and may contain biased recommendations driven by bidding ads, so results require careful vetting.

AI e-commerce is expected to enter a period of broad adoption by 2026, bringing transformative changes and new opportunities for brand operations. Key takeaways for brands are summarized below:

1. Shifts in consumer behavior and competitive dynamics: E-commerce is gradually evolving from traditional shelf-based models and content-driven interest-based commerce to conversational AI e-commerce. User shopping logic has shifted from passive discovery and price-driven purchases to active demand clarification and product matching. The old "price-first" competition logic is being upended, and product quality has become the core of brand competition.

2. New channel opportunities: Major AI tools have already partnered with leading e-commerce platforms. Doubao will soon drive traffic to Douyin Mall, while Qianwen is fully integrated with Taobao. JD and Xiaohongshu have also completed their AI shopping layouts. Brands can secure early positioning by partnering with these AI channels ahead of mass adoption.

3. Adjustments to operational strategy: AI recommendations are primarily based on real user product feedback. To gain higher recommendation weight and more traffic from AI systems, brands need to accumulate a large volume of authentic product experience content and focus on improving product quality.

AI e-commerce is emerging as a new high-growth track for the industry, bringing new opportunities as well as risks that sellers need to navigate. Key takeaways for sellers are summarized below:

1. New opportunities: In AI e-commerce, product discovery is initiated by users, with AI uncovering scenario-based latent demand through conversation. This delivers higher conversion efficiency than traditional passive product discovery. Under the new competitive framework, traffic is no longer won purely through low pricing—small and medium-sized sellers with high-quality products can also earn AI recommendations and access new sources of traffic. All major platforms are currently building out AI e-commerce ecosystems, so early entry allows sellers to capture first-mover advantages.

2. Key risks to watch: AI recommendations rely on authentic user review data, so false promotional claims will lead to incorrect AI recommendations and damage user experience. AI also faces inherent platform information walls that prevent real-time cross-platform price comparison, and there is risk of biased recommendations driven by bidding ranking traps.

3. Recommended actions: Sellers should accumulate authentic product content, focus on improving product quality, proactively adapt to AI recommendation rules, and partner with major platforms' AI channels in advance.

The arrival of the AI e-commerce era brings new insights and business opportunities for factories' product development, design and digital transformation. Key takeaways for factories are summarized below:

1. New direction for product development and design: AI e-commerce can accurately aggregate and organize real user demand, and pass fragmented demand from different use scenarios to the supply side. Factories can leverage AI to access more accurate demand data to guide R&D and design. AI can also uncover scenario-based combined demand, enabling factories to develop bundled product sets that meet users' one-stop shopping needs.

2. Business opportunities: All major platforms are currently improving their AI e-commerce ecosystems. Factories that connect to platforms' AI recommendation systems early can benefit from preferential early-stage traffic allocation, open up entirely new sales channels, and secure more orders.

3. Insights for digital transformation: The core of competition in AI e-commerce is product quality. Factories no longer need to compete on price by compressing margins, and can instead redirect more resources to improving product quality. They also need to advance digital infrastructure to organize authentic product information and user reviews that can be crawled by AI for recommendations, to better adapt to the new ecosystem.

AI e-commerce has become a clear development direction for the e-commerce industry, and industry changes have created new demand and opportunities for service providers. Key takeaways are summarized below:

1. Industry development trends: E-commerce has officially shifted from search-based commerce to conversational AI e-commerce. The core pillars of AI e-commerce are depth of content and breadth of product selection. Going forward, industry demand for AI capability building, authentic content accumulation, and AI recommendation-aligned operational services will grow substantially, creating vast market opportunities.

2. Core pain points for current clients: AI e-commerce still faces several clear pain points in its early stage. First, inter-platform competition has created natural information silos, and AI cannot currently deliver real-time cross-platform price comparison. Second, the accuracy of AI recommendations heavily depends on the timeliness and authenticity of source data, and many merchants lack sufficient accumulated high-quality source content. Third, AI recommendations can become a new form of biased bidding ad trap, damaging user trust and merchant reputation.

3. Opportunities for service providers: Service providers can develop targeted solutions to these pain points, for example helping merchants accumulate authentic user experience content to improve the quality of source data for AI recommendations, and helping merchants build product information systems adapted for AI conversational recommendations, to capture the traffic opportunities of AI e-commerce.

AI e-commerce is set to become the new core competitive advantage for e-commerce platforms, and major players have already started rolling out their offerings. Below is a summary of key takeaways from current industry practices:

1. Product positioning strategy options: The industry currently follows two distinct product approaches. The first, adopted by Qianwen and Doubao, uses conversation to clarify user demand: it identifies users' real needs through conversation before making recommendations, delivering higher accuracy and better user experience. The second, used by JD AI Shopping, retains the traditional shelf-based model, only replacing the search bar with a chat box; it delivers more product options but suffers from low accuracy and poor user experience. New AI layouts are advised to follow the first approach.

2. Core requirements for building successful AI e-commerce: First, platforms need to improve AI's natural language understanding and logical reasoning capabilities. Second, they need to accumulate a sufficiently large library of real user product feedback to use as source data. Both are indispensable.

3. Risks to avoid: Platforms need to quickly resolve current issues including inaccurate recommendations, lack of real-time price comparison, and bidding recommendation traps, to avoid eroding user trust. They also need to adjust platform operational logic, shifting from growth driven by discount promotions to long-term growth built on accurate demand matching, avoid the risk of vicious price competition, and drive healthy industry development.

The e-commerce industry is currently undergoing a new round of iteration. As an entirely new business paradigm, AI e-commerce has entered the piloting and testing phase, and presents many new developments and issues worth studying. Key takeaways for researchers are summarized below:

1. New industry developments: During the 2024 618 shopping festival, all major platforms collectively rolled out AI shopping features, sending a clear industry signal that the AI e-commerce era will officially arrive around 2026. The iteration path of e-commerce has shifted from traditional shelf commerce and interest-based content commerce to conversational AI e-commerce. The core industry logic has shifted from growth driven by low-cost traffic to accurate matching of real user demand. This breaks the "price-first" competitive cycle that has persisted for more than a decade, and drives the industry toward healthy development.

2. New business model: AI e-commerce transforms traditional search-based shopping into conversational shopping, with AI taking on the role of a professional personal buyer. It can uncover users' scenario-based latent demand from non-transactional conversations and complete active discovery and conversion, forming an entirely new business model: user initiates conversation → AI clarifies demand → accurate recommendation → redirect to conversion.

3. Existing open issues: AI e-commerce currently still faces challenges including inter-platform information silos, recommendation accuracy that relies heavily on source data quality, and the potential for new bidding-based recommendation traps. All of these issues warrant further research to explore solutions.