Is AI-powered shopping becoming the core gateway for next-generation consumption? Many have yet to realize that a quiet but comprehensive transformation is reshaping the foundation of the e-commerce industry.
Following industry discussions after the 2026 Spring Festival, Doubao's integration with Douyin E-commerce finally entered a scaled internal testing phase at the end of March. Unlike the previous model of "AI shopping assistant + redirection to third-party platforms," Doubao's "one-command shopping" feature allows users to express their consumption needs, after which it accurately matches product cards from Douyin. Users can click to enter the product detail page and complete their purchase entirely within the Doubao app without switching to Douyin.
The prerequisite is that Doubao and Douyin accounts must be linked. Only after binding their Douyin account can users check corresponding order information within the Doubao app, while product details, the transaction system, and after-sales fulfillment are still handled by Douyin E-commerce.
The core advantage of this model is that ByteDance has streamlined the consumption pathway with minimal effort, quietly shifting the e-commerce transaction entry point from the Douyin app to Doubao's chat interface.
This development did not happen overnight. Tracing its progression reveals a six-month buildup:
- In October 2025, Doubao already had the capability to recommend products from Douyin Mall based on conversation content and enable redirection for purchases, initially covering high-frequency consumption categories like maternal and child products, cosmetics, and home goods. - In November 2025, Doubao launched a product card feature, completing the loop from intelligent Q&A to one-click shopping. - On March 30, 2026, media tests confirmed that users could place orders and make payments directly within Doubao without redirection, marking the completion of an in-app transaction closed-loop.
On one side is Doubao, with 226 million Monthly Active Users (MAU) firmly holding the top spot among domestic native AI apps. On the other is Douyin E-commerce, with a Gross Merchandise Volume (GMV) exceeding 3.43 trillion yuan, solidly ranked third in the industry. The integration of these two ecosystems signifies a deeper exploration by ByteDance into AI commercialization and e-commerce synergy. However, deeply embedding AI commercialization with e-commerce scenarios is a risky move for ByteDance—building trust, technological iteration, ecosystem development, and fulfillment capabilities all present significant hurdles.
**Why is ByteDance in a Hurry to Make its Move?**
Why is ByteDance aggressively pushing Doubao's integration with Douyin E-commerce at a time when user trust in AI shopping is not yet established and the industry still faces numerous bottlenecks?
Largely, this stems from an unavoidable reality facing all AI applications: user numbers are soaring, but commercialization capabilities lag far behind. Internationally, ChatGPT's core revenue still comes from membership subscriptions, with a clear ceiling on commercialization. Domestically, most AI applications are also stuck in the rut of membership fees and enterprise services, unable to achieve revenue growth commensurate with their user scale.
Doubao is in a similar predicament. QuestMobile data shows that as of December 2025, Doubao led domestic native AI apps with 226 million MAU, equivalent to the combined MAU of the 2nd to 5th placed AI applications in China, and is rapidly catching up to the monthly active user scales of national-level apps like Bilibili and Xiaohongshu. During the 2026 CCTV Spring Festival Gala period, Doubao's Daily Active Users (DAU) surged to 145 million.
Faced with continuously rising user metrics, converting this traffic into revenue has become the toughest challenge. After all, the research, development, iteration, computing power procurement, and daily maintenance of large language models all require sustained, massive capital investment. Without finding a scalable, sustainable monetization path, even a vast user base is merely a bottomless pit consuming funds.
A China Merchants Securities research report clearly states: In the era of AI e-commerce, traffic gateways will shift from traditional product shelves to AI agents. User intent-driven advertising will replace keyword search ads, forming a positive feedback loop of "data - model - transaction." The core profitability of this model lies precisely in transaction commissions and intent-based auction advertising revenue.
Based on this, integrating with Douyin E-commerce opens a high-ceiling monetization path for Doubao that aligns with ByteDance's commercial DNA. When users place orders through Doubao, transactions occur within the Douyin E-commerce ecosystem, offering immense potential whether through advertising revenue sharing or transaction commissions.
Of course, ByteDance's ambitions extend far beyond this. A national-level AI application with over 100 million DAU and increasing user stickiness, embedding e-commerce transactions directly into conversational scenarios, has the potential to disrupt not just Douyin E-commerce's existing and new growth but the fundamental rules of the entire e-commerce industry.
Over the last two decades, e-commerce entry points have undergone three major shifts:
1. The first was the search bar era of the PC internet, centered on Taobao and Baidu, where users actively searched for products—a "person finds goods" logic. 2. The second was the live streaming era of the mobile internet, centered on Douyin and Kuaishou, which stimulated consumer demand through content—a "goods find person" logic. 3. The third is the dialog box era of AI, where users merely express their needs, and the AI handles the entire transaction process—an "AI agent" logic.
Correspondingly, each shift in the entry point has rewritten the industry landscape: the search era made Taobao and JD.com; the live streaming era made Douyin and Kuaishou. In the AI era, whoever captures the gateway to conversational transactions will hold the commanding voice of the next era.
Therefore, from the functional evolution of Doubao, it's clear that ByteDance is attempting to position Doubao as the traffic gateway of the AI era. Once users develop the habit of using Doubao for one-command shopping, ordering food, and booking tickets, it transcends being a mere tool-based AI assistant to become the entry point for all internet consumption.
This is also why, even though the feature is still in internal testing, ByteDance has already elevated it to a corporate strategy. ByteDance CEO Liang Rubo explicitly stated at the 2026 all-hands meeting that the company would integrate existing businesses through AI assistants to reach new heights. For now, it appears the "peak" reflected by the group's determination in the short term is Doubao.
**A War of Involved Participation**
In recent years, Douyin E-commerce rode the wave of live streaming's popularity to achieve astonishing GMV growth in e-commerce. However, after reaching a GMV of 3.43 trillion yuan for the full year 2024—a 35% year-on-year increase from 2.54 trillion yuan in 2023—its growth rate has significantly slowed compared to previous doubling-style explosions, indicating that the traffic红利 of live streaming commerce is nearing its peak.
To break through this bottleneck, Douyin E-commerce had already begun a strategic transformation. In 2022, it proposed upgrading from "interest-based e-commerce" to "omnichannel interest-based e-commerce," starting to focus on shelf-based e-commerce by launching the Douyin Mall page tab and promoting the Mall section to a core traffic entry point on the homepage. In the first half of 2024, Douyin E-commerce even prioritized "price competitiveness" as its top annual task, following Pinduoduo, Alibaba, and JD.com in betting on a low-price strategy.
However, Douyin's shelf-based e-commerce mindset has yet to fully form. Users primarily open Douyin for content consumption; shopping is often incidental. Driving generalized users to actively enter the Mall for shopping remains a key focus for Douyin E-commerce.
Now, Doubao's AI shopping scenario presents an opportunity for Douyin E-commerce to break the deadlock:
* On one hand, Doubao's hundreds of millions of users represent traffic fundamentally different from Douyin's content traffic: Douyin users primarily seek entertainment content, with consumption needs being passively stimulated; Doubao users initiate conversations with clear consumption intentions, leading to more precise demand and inherently higher conversion efficiency. * On the other hand, as an independent traffic gateway, Doubao can help Douyin E-commerce break its growth ceiling, forming a positive cycle where "traffic feeds the e-commerce ecosystem, and the ecosystem strengthens AI stickiness." Doubao supplies Douyin E-commerce with highly precise consumer traffic, while Douyin E-commerce enriches Doubao's application scenarios and improves user retention with its mature product pool and supply chain capabilities.
Of course, ByteDance's urgency to go all-in on AI shopping also stems from a more pragmatic consideration: the final battle for the e-commerce industry has begun. Major platforms are accelerating their布局 in the AI shopping arena. Not keeping pace risks being pushed off the competitive landscape.
Looking at recent developments, even before Doubao integrated with Douyin E-commerce, internet giants like Alibaba, JD.com, and Meituan had already staked their claims, leveraging their respective ecosystems for differentiated penetration of AI shopping, culminating in a wave of large-scale implementations during the 2026 Spring Festival period.
* Alibaba's Tongyi Qianwen is the most aggressive player in this race. In January 2026, the Tongyi Qianwen App fully integrated with Alibaba's entire ecosystem, including Taobao Flash Sales, Alipay, Fliggy, and Amap, enabling transactions across shopping, ticket booking, and food ordering. During the Spring Festival, Tongyi Qianwen launched a "3 Billion Yuan Freebies" campaign, completing over 120 million AI-driven orders in 6 days, with related command calls exceeding 4.1 billion. * JD.com began internal testing of its standalone app "JD AI Buy" in late December 2025. Leveraging its self-developed Yanxi LLM, it integrates resources from JD Retail, JD Daojia (food delivery), 7Fresh, JD Health, and others, attempting to build an AI gateway for "e-commerce + lifestyle services." * Meituan started public testing of its standalone AI Agent application "Xiaomei" in September 2025, focusing on voice-based food ordering and restaurant finding, enabling transactions without switching to the Meituan app, deeply embedding itself in instant retail scenarios.
Globally, OpenAI introduced product search and recommendation within ChatGPT in April 2025 (requiring redirection to external platforms). On September 29, it launched Instant Checkout, achieving an in-app, no-redirection transaction closed-loop (initially with partners like Etsy and Shopify). In October 2025, Walmart partnered with OpenAI to enable a shopping gateway within ChatGPT via a Sparky embed. This partnership ended in March 2026, with Walmart shifting towards a self-developed multi-platform embedding solution.
A consensus is gradually forming across the industry: AI shopping is the core gateway for the next generation of consumption. Consequently, ByteDance appears to have been involuntarily drawn into an arms race centered on AI shopping. Participation is necessary to remain relevant; failure to keep up could mean losing the hard-won gateway advantage as users are gradually siphoned off by competitors.
**Hurdles Facing ByteDance**
Even though ByteDance holds a strong hand with its traffic assets and attempts to replicate its success stories from short video and e-commerce, the path to maturity for AI shopping is fraught with challenges related to trust, technology, ecosystem, and fulfillment.
The most significant short-term problem is: Why should users trust AI recommendations?
The core premise of users' willingness to shop online is the ability to autonomously build a decision-making system through actions like checking product details, browsing buyer reviews, comparing store ratings, and cross-store price comparisons. These seemingly tedious actions are key to user decision-making. AI shopping essentially hands this entire decision-making process over to the AI, causing users to lose control and passively accept recommended results, inevitably conceding their autonomy in choice.
Although AI is advancing rapidly, the current e-commerce ecosystem is already seeing AI-generated buyer showcases, excessively beautified product images, and hollow feedback reviews. If AI recommendation logic becomes influenced by advertising budgets or brand partnerships—if the platform uses AI to generate fake product content and polish marketing materials while letting the same AI wield decisive power over product recommendations and purchase decisions—the foundation of trust essential for e-commerce could inevitably crumble.
Currently, Doubao primarily outputs results. How can the objectivity and authority of the underlying recommendation logic and decision-making basis be guaranteed? Even top live streamers frequently encounter issues like false advertising and product mismatches. If AI shopping leads to inaccurate recommendations or quality control failures, it could severely damage Doubao's credibility and Douyin E-commerce's brand reputation.
Of course, user distrust also largely stems from the inherent conflict between AI hallucinations and real-world consumption experiences. The AI shopping promoted by major tech companies relies on the technical capabilities of their LLMs. However, even the industry's most advanced models cannot solve the innate problem of AI hallucinations. Most currently deployed AI shopping features intermittently suffer from incorrect product parameters, inaccurate prices, recommendations mismatched with user needs, and delayed updates on promotional information. These native AI hallucinations can single-handedly set e-commerce back, as e-commerce transactions require extremely high information accuracy. Inaccuracies in price, specs, inventory, or promotions directly impact the user shopping experience and can lead to numerous after-sales disputes.
Furthermore, the experience of "browsing" and the satisfaction of finding the best deal are integral parts of consumption. For many, shopping provides not just utility but also emotional value. Simplifying consumption to a "one-command order" risks turning the enjoyable act of shopping into a cold, transactional instruction, stripping away the emotional engagement of the process.
Moreover, AI shopping can only effectively cover standardized, purpose-driven essential consumption scenarios. For non-standardized products, high-ticket items, or scenarios requiring deep experiential evaluation, AI can never fully replace user autonomy in decision-making.
Another unavoidable premise is that true interconnectivity within the current mobile internet remains largely utopian. What is termed "AI shopping" is often merely a pseudo-"one-stop shopping." Essentially, Doubao's product supply relies on Douyin E-commerce; Tongyi Qianwen only accesses Alibaba's ecosystem; JD AI Buy integrates only JD.com's own categories—all are confined to their respective platform's closed loops. However, what users truly desire from AI shopping is a solution that addresses consumption needs across the entire internet, not just within a single, closed corporate ecosystem. Doesn't this make the promoted "conversation as transaction" currently a somewhat false proposition? It often merely consolidates operations previously scattered across different apps into a single dialog box, while the underlying product pools and fulfillment systems remain siloed. To compare prices across platforms or purchase items not listed on a specific platform, users still need to switch to other apps.
Simultaneously, closed ecosystems further exacerbate data silos. The accuracy of AI recommendations heavily depends on vast amounts of user data across all consumption scenarios. Doubao can only access consumption data within the Douyin E-commerce ecosystem, unable to tap into full-scenario consumer behavior, inherently limiting the ceiling of its recommendation capabilities.
Finally, the ultimate competition in e-commerce extends beyond front-end traffic and conversion to back-end fulfillment and service. On the surface, Doubao enables "one-command ordering," but this is limited to the ordering step itself. Product details, the transaction system, and the after-sales system all belong to Douyin E-commerce. If issues like product quality problems, shipping delays, or after-sales disputes arise, users still need to communicate across applications; a truly seamless after-sales and fulfillment closed-loop has not been fully established.
The core issue behind this situation is ByteDance's inherent weakness in fulfillment systems. The final moat in the e-commerce industry lies in supply chain and fulfillment capability: JD.com built its defensive wall with self-operated logistics; Meituan secured its top spot in local services with a nationwide instant delivery network; Pinduoduo achieved its low-price advantage through extreme supply chain management. In contrast, ByteDance's commercial DNA is rooted in extreme ROI orientation and operational efficiency priority, often showing impatience towards long-cycle, capital-intensive heavy-asset布局 like logistics fulfillment. Consequently, Douyin E-commerce, to this day, has not established its own logistics system, relying entirely on third parties, resulting in a gap in delivery efficiency and service experience compared to players like JD.com with self-built networks.
A deeper industry concern is whether the full-scale implementation of AI shopping will reconstruct merchants' operational logic and, like a domino effect, further impact the platform's merchant ecosystem. Data from the "2025 Live Streaming E-commerce Development White Paper" shows that the concentration of traffic and sales share among top live streamers is continuously declining, with mid-tier streamers and small-to-medium merchants gradually becoming the market's backbone. However, in the AI e-commerce era, major brand merchants clearly have more substantial budgets to adapt to AI recommendation algorithms and capture traffic红利. Small and medium-sized merchants, meanwhile, face new costs associated with learning AI tools, data services, and rule adaptation, potentially seeing their生存空间 squeezed further.
Even if Douyin E-commerce continuously optimizes its algorithm distribution mechanisms and opens its AI capabilities to merchants, aiming to help them reduce costs and increase efficiency across customer service, content creation, budget management, and precision marketing, a fundamental矛盾 remains: when AI holds absolute sway over product recommendations, merchants' focus will inevitably shift from refining products and improving service to adapting to AI recommendation rules and competing for AI-driven traffic倾斜, potentially leading the e-commerce ecosystem into a new cycle of intense internal competition.
Comments