Kuaishou Launches AI Shopping Assistant in Search Results as China's AI Shopping Guide War Intensifies

What Kuaishou Embedded in Its Search Result Pages

Kuaishou Quietly Launches AI Shopping Assistant, Escalating the AI Shopping Guide War Among China's E-Commerce Giants

Kuaishou E-commerce quietly launched an AI shopping assistant on July 20, embedding the entry point directly into product search result pages.

finance.biggo.com

There was no press release and no launch event. On July 20, Kuaishou E-commerce placed an entry point for its "AI Shopping Assistant" in the bottom-right corner of product search result pages in its app and quietly switched it on. Users can invoke the AI immediately after searching. Describing what they want in natural language brings up a horizontal row of matching product cards, and tapping the cart icon on a card completes the add-to-cart step without opening a product detail page. The design compresses the traditional chain of search, browse, compare, and purchase about as far as it can go.

In practice, the assistant is relentlessly utilitarian. Typing "buy an iPhone 17E under 5,000 yuan" makes the AI instantly filter options across price tiers, such as a new unit at 4,499 yuan with 12-month interest-free installments and a subsidized near-new unit at 4,299 yuan after coupons. Users pick the color and storage and add to cart without ever visiting a detail page.

For comparison-heavy categories like consumer electronics and home appliances, the support goes further. Telling the AI you are torn between the OnePlus Ace 6T and the iPhone 17E produces an on-the-spot comparison table covering core specs, effective prices, sales track record, reputation, and pros and cons. Even vague requests like "affordable kitchen finds" or "sun protection gear for a trip" return candidates sorted by sales volume and review scores. If a previously browsed product drops in price or its warranty terms change, a notification arrives, and instead of reading thousands of reviews, users can ask the AI to summarize the main points of positive and negative feedback along with common pitfalls.

Combining AI and e-commerce is not new territory for Kuaishou. Over the past year, its AI has been concentrated in merchant-facing back-end tools such as image-to-video generation and livestream clipping, which supported an average of 1.5 million videos produced per day during the Singles' Day season and drove average daily GMV of 53 million yuan, up 140.9% year on year. Those were back-office capabilities invisible to ordinary users. This is the first time Kuaishou has put AI directly in front of consumers.

The product philosophy is also distinctive. It is neither a multi-turn dialogue that teases out preferences nor a generator of lengthy purchasing advice; it is single-mindedly about removing steps. If other platforms position their AI shopping guides as "your consultant," Kuaishou has taken the "saving you steps" efficiency route, carrying over the style of its "laotie" (loyal buddy) economy rooted in users from China's lower-tier cities.

Taobao, Douyin, and JD.com: Three Different AI Shopping Guide Strategies

Zooming out to the whole industry, Kuaishou's move is not a first-mover play. China's e-commerce giants are already running AI shopping guides, each in a different configuration.

Alibaba laid its groundwork earliest. It committed fully to AI commerce at the end of 2023, starting with merchant-side tooling such as store operation assistants, and then for Singles' Day 2025 it launched six AI shopping guide apps in one go, headlined by "AI Universal Search." The company completed a semantic restructuring of its 2-billion-product database and reported a 20% improvement in search relevance. Even so, the full purchase loop connecting its standalone AI apps outside the Taobao ecosystem with on-platform transactions is still a work in progress.

Douyin chose the "blooming outside the wall" route. In October 2025 it connected its AI app Doubao to Douyin Mall, so that asking Doubao about beauty or electronics products surfaces product cards with prices and sales figures, and a tap jumps to Douyin to place the order. It has since been testing the Doubao Phone Assistant, which handles cross-app price comparison and ordering, signaling ambitions beyond its own ecosystem.

Just five days before Kuaishou's launch, on July 15, JD.com teamed up with Tencent. JD.com's AI agent formally connected with the mini-program ecosystem of Tencent's AI app Yuanbao. Asking Yuanbao "which refrigerator under 2,000 yuan is the quietest" brings up purchasing advice alongside JD.com product cards, and the order completes inside the JD.com mini-program, with logistics and after-sales handled by JD.com and no app switching involved. JD.com has also connected with device-side AI assistants such as Huawei's Xiaoyi and OPPO's Xiaobu through the A2A (agent-to-agent) model, effectively covering all mainstream AI entry points in a "cast a wide net" strategy. If Douyin is the vertically integrated model that funnels its own AI traffic to its own mall, JD.com treats AI purely as a traffic pipeline and monetizes through its supply chain and fulfillment capabilities.

Why the Latecomer Is Moving Now: A Growth Plateau

Why this timing? The answer is in the financials.

Kuaishou's e-commerce GMV reached 1.6 trillion yuan in 2025, up 15% year on year. The number looks respectable on its own, but compared with 78% growth in 2021 it represents a deceleration of 63 percentage points in five years. On the user side, average daily active users in 2025 were 410 million, up only 2.76%, and monthly active buyers in the first quarter of 2025 stood at 135 million, leaving a visible gap with Douyin E-commerce. Industry estimates put Douyin's 2025 e-commerce GMV at roughly 4.4 trillion yuan, about 2.75 times Kuaishou's.

The revenue picture offers little slack either. First-quarter 2026 revenue was 33.72 billion yuan, up just 3.4% year on year, and securities analysts forecast full-year growth of around 4.5%. The video generation AI Kling posted more than 650 million yuan in revenue for the quarter, but AI investment as a whole has not reached the payback phase. With traffic dividends thinning and shelf-based e-commerce yielding no clear differentiation, lifting the conversion rate of its core e-commerce business with AI is one of the few moves that improves internal economics and gives capital markets a growth story at the same time.

The Investor View: AI Inference Cost Pessimism Is Overdone

In parallel with the platforms' moves, capital market sentiment is starting to shift.

In a recent report, Bernstein analyst Robin Zhu argued that some AI cost narratives circulating in the market, such as the idea that platforms must subsidize inference costs indefinitely, are "grossly inaccurate and excessively pessimistic." According to the report, chatbot-style conversations typically consume only a few hundred tokens, while agentic transactions, where an agent carries a purchase through to completion, average about 50,000 tokens. In other words, inference costs only explode when agent-driven transaction volumes and GMV genuinely take off, and in that scenario revenue would be rising at the same time. Bernstein set a price target of HK$780 for Tencent, implying roughly 69% upside, and identified the trillions of yuan in GMV flowing through WeChat mini-programs as the core opportunity for Tencent's AI monetization.

That said, operational assessments remain cautious. As the source article notes, AI shopping guides across the industry are still in a "nice-to-have" phase. They lift conversion rates by a few percentage points and shave a few steps off the purchase journey, but they are far from restructuring the industry, and whether "telling an AI one sentence and buying" becomes a user habit will take time to verify.

Implications for E-Commerce Operators: Preparing for Two Funnels

This competition unfolding in China is not a distant spectacle for e-commerce operators in other markets, including Japan. ChatGPT's shopping features and Google's AI Mode point in the same direction: the entry point for product discovery is moving from the search box to conversational AI worldwide.

Sorting through the Chinese giants' moves reveals two funnel types. One embeds AI inside the operator's own app, as Kuaishou and Taobao do. The other hands product data and fulfillment capability to external AI assistants, as JD.com does. Few operators have the resources to build their own AI shopping guide, so for most, the realistic first step is the latter: structuring product data so external AI agents can correctly discover, compare, and recommend their products.

Kuaishou's feature list doubles as a checklist. If AI auto-generates comparison tables, products without structured specs, prices, and warranty terms never make it into the comparison at all. If review summarization becomes standard, the volume and quality of reviews will shape recommendation outcomes. If price-drop alerts spread, the freshness of price and inventory data will translate directly into customer experience quality. The rise of AI shopping guides is quietly raising the value of unglamorous product data work.

Conclusion

With the dividends of livestream commerce exhausted and the shelf-based e-commerce landscape largely settled, China's e-commerce giants have chosen "the efficiency of matching people with products" as the next axis of competition. Kuaishou's AI shopping assistant arrives late, but it enters the race with the shortest-path design available: an entry point wired directly into search result pages. Alibaba holds its supply chain, ByteDance its traffic, Tencent and JD.com their alliance, and Kuaishou its base of loyal users; each plays a different hand. Where AI agents will stand in the flow of purchasing is still being mapped, and this contest is only in its opening moves.