Qianfeng Interview with Wang Yueneng of Baijiahui: The Modal Evolution of B2B Cross-border Trade Moves from 'Product Going Global' to 'Capability Going Global'

【Ebrun Original】Wang Yueneng is the Chairman of Hangzhou Baijiahui Foreign Trade Group Co., Ltd. Baijiahui has been deeply engaged in foreign trade for thirty years, adopting a trinity model of 'traditional foreign trade + internet + overseas localized operations.' Based on its self-developed global digital supply chain and B2B trading platform 'China Foreign Trade Expert,' it provides full-chain services from inquiry and contracting to customs clearance, logistics, and foreign exchange settlement. In this Qianfeng Award interview, Wang Yueneng shared Baijiahui's practices and judgments in the B2B cross-border trade direction, focusing on the modal evolution brought by AI and going global.

Ebrun: This year's Qianfeng Award takes 'The Modal Evolution Brought by AI and Going Global' as its core theme, focusing on the deep reconstruction of industrial capabilities. Combining B2B cross-border trade and comprehensive foreign trade services, how do you understand this round of industrial modal evolution jointly driven by AI and globalization? What core strategic layouts does Baijiahui have in AI capability building and global expansion? Could you specifically introduce the company's positioning and business model?

Wang Yueneng: The essence of this round of industrial modal evolution is a historic leap for China's foreign trade from 'product going global' to 'capability going global,' and from 'world factory' to 'global supply chain organizer.' It occurs at the intersection of three certain trends: First, China firmly advances high-level opening-up and institutional opening-up. New foreign trade formats and models (cross-border e-commerce, overseas warehouses, market procurement) have been incorporated into national top-level deployment, opening institutional space for B2B cross-border trade. Second, the Belt and Road Initiative has entered a new stage of high-quality development. Industrialization and infrastructure needs in Global South countries serve as the annual bellwether for the TO B sector. The '2026 8th Ebrun Industrial Internet Annual Conference and First Industrial Intelligent Agent Conference' will be held in Beijing on November 20, 2026. The theme is set as 'The Modal Evolution Brought by AI and Going Global,' focusing on deep deliberation of frontier topics such as the implementation of industrial intelligent agents, value reconstruction of AI supply chains, data elements and the TOKEN economy, and the new ecosystem of industrial going global. The conference will gather academic authorities, benchmark central state-owned enterprises, industrial leaders, investment institutions, and service ecosystem partners, pooling core industrial forces, deconstructing industrial digital-intelligent transformation paths, and sharing frontier practical experience. Concurrently, the annual industrial internet report and benchmark cases will be released, and a decision-makers closed-door meeting and the 'Qianfeng Award' ceremony will be held. Demand continues to be unleashed, making it the most certain incremental market in cross-border trade. Third, new quality productive forces represented by artificial intelligence are reconstructing the 'efficiency gap' and 'trust gap' across the entire foreign trade chain into calculable and replicable capabilities.

Baijiahui has been deeply engaged in foreign trade for thirty years. Its model is a trinity of 'traditional foreign trade + internet + overseas localized operations': based on the self-developed global digital supply chain + B2B trading platform 'China Foreign Trade Expert,' it deeply embeds into transaction and fulfillment links, integrates domestic industrial belt factories and overseas buyer resources, and provides full-chain services from inquiry and contracting to customs clearance, logistics, and foreign exchange settlement. In the AI direction, it has completed full local deployment of AI capabilities and built a data governance system spanning the entire business chain, allowing AI to truly grow within processes. In the globalization direction, with overseas operation centers and public overseas warehouses as pivots, it has deployed local teams and channel networks in Belt and Road countries and key Global South markets, achieving 'direct connection between Chinese products and overseas terminals.' AI improves matching and fulfillment efficiency, while localization builds trust and delivery capability—together they support the platform's value closed loop of 'trade certainty.' This is the most concrete manifestation of what we understand as 'modal evolution' at Baijiahui.

Ebrun: You mentioned that Baijiahui's model is a trinity of 'traditional foreign trade + internet + overseas localized operations.' How exactly do these three synergize? Why not simply 'traditional foreign trade + internet'?

Wang Yueneng: The core of the trinity is that traditional foreign trade provides thirty years of customer resources, channel accumulation, and fulfillment experience; the internet provides digital platforms and AI capabilities; and overseas localized operations provide local teams, channel networks, and fulfillment nodes. The three are not additive but interlocking: AI improves matching and fulfillment efficiency, localization builds trust and delivery capability, and both jointly support the platform's value closed loop of 'trade certainty.' If there were only 'traditional foreign trade + internet' without localization, it would merely move offline processes online and could not solve the trust and delivery problems in cross-border transactions.

Ebrun: Currently, in which core business links has your company implemented AI/large model technologies? What distinctive AI or digital-intelligent competitiveness system has been formed?

Wang Yueneng: Our AI implementation follows one principle: no concept stacking, only hitting pain points. Foreign trade is a typical industry with 'long chains, multiple roles, and heavy trust.' The value of AI lies not in showing off technology but in deeply and thoroughly addressing high-frequency, repetitive links with data accumulation. Currently, capabilities are mainly built around four major scenarios:

First, intelligent matching and product selection. Based on supply-demand and fulfillment data accumulated over thirty years, overseas procurement needs are precisely matched with domestic industrial belt capacity, greatly compressing the 'find goods—compare goods—order goods' cycle.

Second, multilingual intelligent agent customer service. Covering key market languages, it enables cross-language real-time communication and removes the natural barrier that 'language is a threshold.'

Third, AI document recognition and intelligent review. Targeting the pain points of foreign trade documents being 'many in type, mixed in format, and zero-tolerance for errors,' it achieves automatic recognition, element extraction, and logical verification, freeing people from the most tedious entry and checking.

Fourth, internal process intelligence. It assists decision-making in risk control, customer operations, fulfillment tracking, and other links, enabling 'those who can hear the guns' to call AI capabilities in real time.

Ebrun: Could you share a most representative 'AI + value chain' innovation practice case? What common industry pain points did it solve, and what perceptible, quantifiable value did it create?

Wang Yueneng: The most representative is AI document recognition and intelligent review. Foreign trade document processing is a typical 'high-frequency, trivial, zero-tolerance' link: an order from contract, invoice, packing list to customs declaration documents involves a large number of files in inconsistent formats. Under the traditional manual model, it is labor-intensive and error-prone, and a single document error can cause real financial losses such as customs clearance delays and letter of credit discrepancy deductions. After applying large models to automatic document recognition, element extraction, and logical verification, business personnel shifted to handling exceptions and customer service, and the efficiency and accuracy of document processing improved significantly, with the stability of customers' customs clearance timeliness rising accordingly.

The deeper value lies in: AI does not necessarily need to do flashy front-end work; solidly doing the industry's most 'dirty and hard' links can instead create value perceivable by both upstream and downstream—factories run fewer errands, buyers wait less, and customs clearance gets stuck less.

Ebrun: You mentioned that Baijiahui adopts full local deployment of AI, with the principle of 'configure by scenario, sufficient and scalable.' How exactly do you control computing power costs? What can you share about TOKEN consumption and computing power costs?

Wang Yueneng: Regarding computing power costs, we adopt full local deployment of AI, and the investment falls within infrastructure construction. The principle is 'configure by scenario, sufficient and scalable': costs are controlled through scenario priority ranking and tiered model invocation (large models for complex tasks, lightweight models for simple tasks), ensuring that input-output is measurable. Specific TOKEN and computing power costs are not disclosed for now. This also aligns with the orientation of 'cost reduction and efficiency improvement, pragmatic implementation' in the national 'AI+' action.

Ebrun: In advancing large-scale AI implementation, what supporting role do data elements play? Does Baijiahui focus more on deepening its own industrial data or on promoting cross-entity, cross-industry data collaboration? What are the core challenges?

Wang Yueneng: In the foreign trade industry, the core of data competitiveness lies not in 'volume' but in 'chain completeness.' Baijiahui has been deeply engaged in transactions and fulfillment for thirty years and has accumulated full-chain data from inquiry, matching, and contracting to customs clearance, logistics, and foreign exchange settlement. This is precisely the prerequisite for AI to truly generate business value.

We focus on deepening our own industrial data: first build a data governance system, standardize and structure historical business data, and then overlay full local deployment of AI so that models are trained and optimized on real business data. The biggest challenge is governing old data—data accumulated over thirty years has diverse formats and uneven quality, and we have spent great effort on cleaning and structuring. But looking back, this step determines whether AI applications 'remain in demonstrations' or 'grow within the business.' At a time when 'Data Elements ×' and 'trusted data spaces' are accelerating implementation, this data asset spanning the entire trade chain is precisely our strategic foundation for the future.

Ebrun: How do you view industrial intelligent agents rewriting B2B? Is there a long-term development need for vertical large models? Where lies the AI moat of industrial internet companies?

Wang Yueneng: We judge that AI evolution in the B2B field will go through three stages: toolization (efficiency improvement) → processization (reconstructing workflows) → intelligentization (autonomous execution in some links). Currently most are in the first and second stages. In the future, intelligent agents will gradually take over standardized links such as inquiry response, document processing, and fulfillment tracking, while human roles shift to handling exceptions and building trust.

Regarding the moat, general large model capabilities will eventually homogenize. The real barrier for industrial internet companies is the combination of 'industry knowledge + business data + fulfillment capability.' Large models can understand language, but they do not understand the customs clearance practices of a particular overseas market, do not know the quality inspection standards of a certain category, and cannot be responsible for transaction outcomes. The know-how and data deposited from these 'dirty and hard' tasks, plus our trinity service capability of deep fulfillment, are the parts that AI cannot easily replicate. Vertical models need not pursue being large and comprehensive; the key is to refine and deepen industry-specific scenarios—this also echoes the national direction of encouraging pragmatic development of 'vertical large models and industry large models.'

Ebrun: Regarding overseas layout, which overseas core markets has Baijiahui currently deployed? What key localization adjustments have been made?

Wang Yueneng: Our core markets are concentrated in emerging markets along the Belt and Road and in the Global South. These markets are in a period of accelerated industrialization and urbanization, and digital penetration still has huge room for improvement. This is also where our thirty years of accumulated customer resources and channel advantages lie. This layout is highly consistent with the national direction of 'high-quality Belt and Road cooperation' and the 'Global Development Initiative.'

In localization, we have done three things: First, team localization. We set up local teams in key markets and recruit local marketing, customer service, and channel personnel to achieve 'local people serving local markets.' Second, channel localization. We establish long-term stable cooperation with local leading wholesalers and distribution channels, adapting to local business culture and transaction habits. Third, service localization. We adjust fulfillment plans around local payment habits, logistics conditions, and after-sales expectations. In essence, B2B cross-border business is trust business, and localization turns 'distant suppliers' into 'partners nearby.' At a time when anti-globalization undercurrents and supply chain restructuring coexist, the depth of localization is the capacity to resist risk.

Ebrun: In building deep value chains in overseas markets, what model has Baijiahui formed? You mentioned 'from trade to trust, from trust to ecosystem.' How exactly should this be understood? What are the core capabilities?

Wang Yueneng: Our overseas value chain building follows the path of 'from trade to trust, from trust to ecosystem,' forming three layers of capability: First layer, localized service network. Relying on overseas operation centers, we extend the service radius to customers' side, achieving response in local languages, local habits, and local time zones. Second layer, localized fulfillment capability. Using public overseas warehouses as nodes to integrate logistics, we shorten the delivery chain so that overseas buyers gain an experience close to local procurement. Third layer, reverse supply chain integration. Relying on domestic industrial belt resources, we provide overseas partners with one-stop support in product selection, quality control, and flexible supply, turning 'Chinese supply' into reliable certainty capability.

For overseas partners, we create three core values: certainty (quality and delivery are predictable), efficiency (one-stop replacement of multi-party coordination), and risk sharing (someone is responsible for problems to the end). This is precisely the microscopic footnote of our move from 'selling products' to 'exporting trade service capability.'

Ebrun: In the process of overseas expansion, what core support do AI and digital capabilities provide for cross-border business?

Wang Yueneng: The biggest friction costs in cross-border business come from three places: language, trust, and chain complexity. Our AI and digital capabilities are built precisely around these three points. In global supply chain collaboration, the digital platform connects the entire chain of product selection, quality inspection, consolidation, and going global, achieving full-process logistics visibility. In cross-border fulfillment, AI document recognition and multilingual intelligent customer service significantly compress communication and document time and reduce error rework. In overseas customer operations, customer profiles based on full-chain data help us understand procurement preferences in different markets, turning passive response into proactive service.

Ebrun: You mentioned an insight: digitalization allows 'small factories to also do big foreign trade.' How should this judgment be understood? What role does the platform play in it?

Wang Yueneng: Many industrial belt factories have strong product power but lack the ability to take international orders. The platform 'translates away' thresholds such as language, compliance, and logistics, allowing them to participate in global competition on equal terms. This is precisely the unique value of platform enterprises under the 'dual circulation' pattern: connecting domestic industrial belts with the global market and smoothing domestic and international dual circulation.

Ebrun: Could you introduce the iteration of Baijiahui's profit model? How do you view the evolution trend of profit models for industrial internet companies? How can TOKEN payment become a standard industry revenue model?

Wang Yueneng: Our profit model has undergone clear iteration: early on, it was mainly based on margins from traditional foreign trade services. With the construction of the 'China Foreign Trade Expert' platform, it gradually shifted to a diversified structure of 'platform service fees + supply chain service value-added + transaction services'—from earning money from information gaps to earning money from efficiency improvement and service value. This corresponds to the evolution trend of the industrial internet from 'channel-based charging' to 'value-based charging.'

As for whether TOKEN payment can become an industry standard, the key is whether AI creates perceptible, measurable incremental value. If intelligent agents truly replace manual links and bring certain cost reduction and efficiency improvement, customers will be willing to pay for results. But pure Token-based billing may not fit all scenarios; 'pay-for-performance' and 'share-of-savings' may better fit the B2B mindset. Our exploration direction is to embed AI capabilities into the trinity service system and make them part of the overall solution value rather than an isolated billing item—this both serves customer interests and reflects the platform's commitment to 'value co-creation.'

Ebrun: To coordinate AI capability implementation, domestic value chain upgrading, and overseas business expansion, what key adjustments has Baijiahui made in organizational structure or governance mechanisms? You mentioned 'let those who can hear the guns call AI capabilities.' How exactly is this implemented?

Wang Yueneng: Our organizational adjustments unfold around 'let those who can hear the guns call AI capabilities': First, platformize AI capabilities. We consolidate AI and data capabilities into an internal shared middle platform, which business lines can call as needed, avoiding duplicate construction and unifying governance standards. Second, integrate business and technology teams. In core scenarios, business backbone personnel and technical personnel form special teams; business defines problems and technology implements and accepts them, shortening the translation chain from 'business language' to 'technical language.' Third, establish domestic and overseas coordination mechanisms. Overseas local teams handle market front-end and customer relationships, while domestic teams carry the supply chain and fulfillment hub, achieving cross-time-zone collaboration and frontline authorization through the digital platform.

The key to ensuring implementation is 'aligned goals + appropriate authorization': the headquarters controls the strategic direction, while execution flexibility is given to the frontline, with data dashboards making the process transparent—neither deviating from the course nor losing the flexibility for innovation. This mechanism itself is an inevitable requirement of 'new quality productive forces' for organizational form.

Ebrun: In your view, what are the biggest internal bottlenecks and external obstacles currently constraining large-scale AI implementation and deep global expansion for enterprises?

Wang Yueneng: The biggest internal bottleneck is composite talent: people who understand foreign trade, AI, and overseas markets are extremely scarce. AI implementation is often difficult not because of technology but because of the 'translation' between business and technology. We alleviate this by having business personnel deeply participate in AI application design.

The biggest external obstacle is the complexity and uncertainty of the cross-border environment: regulatory regimes, data compliance, and trade rules differ greatly across countries and change dynamically. Our response is to build compliance as infrastructure—at the data level, full local deployment of AI meets security and compliance requirements; at the trade, tax, and financial levels, we establish a compliance system covering the entire process. We always believe that compliance is not a cost but the ticket and moat for long-term going global; the more high-level opening-up advances, the more high-level compliance capabilities are needed to match it.

Ebrun: Looking toward 2026 and the next two years, what key investment plans does Baijiahui have in AI capability iteration and global layout? What key changes do you predict this will bring to the industry competitive landscape?

Wang Yueneng: In the next two years, key investment will concentrate in three directions: In AI, upgrade from 'scenario applications' to 'intelligent agent collaboration,' allowing intelligent agents to gradually and autonomously take over standardized links such as inquiry, documents, and fulfillment tracking, and continue to deepen data assetization. In globalization, deepen localization in core markets, improve the overseas fulfillment network, and replicate validated models to more emerging markets. In the ecosystem, connect more industrial belt factories and overseas channels through 'China Foreign Trade Expert' to expand network effects.

Regarding the industry landscape, we have two predictions: First, differentiation will accelerate—the gap between platforms with comprehensive capabilities in 'data + fulfillment + localization' and pure information matching platforms will widen rapidly, and AI will amplify this gap. Second, the competitive dimension will shift from traffic to trust and delivery—after the online traffic dividend peaks, whoever deepens fulfillment certainty and localization will take the initiative. Overall, this will push B2B cross-border trade from 'channel competition' into a new stage of 'capability competition.'

Ebrun: Have opportunities emerged in industrial goods e-commerce? In what way do you predict it will grow rapidly?

Wang Yueneng: The opportunity is clear, driven by three major forces: First, continued demand for overseas infrastructure drives stable exports of construction machinery, building materials, hardware tools, and other goods. Second, manufacturing upgrading in emerging markets—industrialization in Southeast Asia, the Middle East, and Latin America generates strong demand for intermediate goods and equipment. Third, industrial goods B2B has lower digital penetration and relies more on intermediaries, so the room for transformation is actually greater—it is precisely a 'digital lowland.'

In terms of growth mode, we predict it will not be a pure online traffic explosion but gradual growth in a hybrid model of 'online matching + offline fulfillment + local services': industrial goods have long decision cycles, high unit values, and heavy pre-sales and after-sales, and customers value trust and delivery certainty more. The first to break out will be platforms that can provide comprehensive solutions of 'product selection consulting + quality assurance + localized after-sales + supply chain finance'—this is consistent with our trinity direction of making services deeper and heavier. For platforms deeply engaged in industrial belts, industrial goods are one of the most certain incremental tracks in the coming years.

Ebrun: If you were to give one core piece of advice to peers who have just started and hope to build both AI competitiveness and globalization capabilities, what would it be?

Wang Yueneng: First solidify the foundation of data and business, then talk about AI and globalization. AI does not generate value simply by connecting to a large model API, and globalization is not layout simply by registering an overseas account. Our experience is: first digitize business processes and make data governance solid, so that AI has soil to grow; first deepen localization in a few core markets and get the fulfillment closed loop running, then talk about scale replication. Slow is fast—companies that lay a solid foundation can truly benefit in the dual transformation of AI and globalization, and can also bear the role of participants in 'high-level opening-up' and connectors of 'dual circulation.'

As the annual bellwether in the TO B sector, the '2026 8th Ebrun Industrial Internet Annual Conference and First Industrial Intelligent Agent Conference' will be held in Beijing on November 20, 2026. The theme is set as 'The Modal Evolution Brought by AI and Going Global,' focusing on deep deliberation of frontier topics such as the implementation of industrial intelligent agents, value reconstruction of AI supply chains, data elements and the TOKEN economy, and the new ecosystem of industrial going global. The conference will gather academic authorities, benchmark central state-owned enterprises, industrial leaders, investment institutions, and service ecosystem partners, pooling core industrial forces, deconstructing industrial digital-intelligent transformation paths, and sharing frontier practical experience. Concurrently, the annual industrial internet report and benchmark cases will be released, and a decision-makers closed-door meeting and the 'Qianfeng Award' ceremony will be held.

This article was first published on the Ebrun official website.