广告
加载中

阿里国际站旗下Accio全球用户增长逾30倍 开源首个真实电商Agent基准测试

亿邦动力 2026-09-10 11:25
亿邦动力 2026/09/10 11:25

邦小白快读

EN
全文速览

本文核心信息是阿里国际站在CoCreate 2026海外买家大会上发布了AI电商相关成果,干货内容如下:

1. 阿里国际站总裁张阔提出B2B贸易的未来是A2A,也就是Agent对Agent做生意,目前该模式已经落地,海外采购商已经开始用自然语言描述需求,由AI完成找趋势、选产品、筛供应商、谈条款、管订单全流程。相关数据显示,AI多模态搜索流量增长超132%,旗下AI工作台Accio全球用户一年增长逾30倍,现有1000万中小企业在用。

2. 普通做跨境生意的用户可获得的实操便利是,Accio已经打通了国内外所有主流电商平台,包括国内淘宝、天猫等和海外Amazon、TikTok等,真正实现多平台生意统一管理,而且Accio的AI token成本比通用Agent低50%,中小企业也能用得起日常使用的AI工具。

对于出海品牌商而言,本文透露了跨境B2B领域的消费新趋势和商业机会,干货内容如下:

1. 海外采购端已经发生结构性变化,采购商不再使用关键词搜索,转而用自然语言描述需求,AI多模态搜索流量已经增长超132%,说明用户已经普遍适配AI采购模式,品牌需要调整流量运营、产品展示策略适配新的交互方式。

2. 品牌可以借助新工具获得多渠道运营优势,Accio现有1000万中小企业用户,已经打通国内外所有主流电商平台,品牌可以实现多平台生意统一管理,降低运营管理成本,而且Accio的AI token成本比通用Agent低50%,可以日常使用,不会带来过高成本压力。

3. 未来B2B贸易会走向A2A模式,品牌可以提前布局AI自动交易,抢先占据新赛道的优势,跟上贸易形态变化的新趋势。

对于做跨境电商的卖家而言,本文透露出新的增长机会和运营方向,干货内容如下:

1. 市场需求已经发生变化,海外采购商已经转变采购习惯,从传统关键词搜索转向自然语言描述需求,AI多模态搜索流量增长超132%,说明AI化采购已经成为新的流量趋势,卖家需要及时调整运营逻辑适配新规则,避免错失流量。

2. 卖家可以获得实用工具支持,Accio成立一年多全球用户增长超30倍,现有1000万中小企业使用,已经打通国内外所有主流电商平台,卖家可以实现多平台生意统一管理,不用切换多个后台,而且Accio的token成本比通用Agent低50%,中小卖家也能负担得起日常使用。

3. 未来B2B贸易会走向A2A模式,提前接入AI工具的卖家能抢先获得匹配优势,拿到更多订单,要抓住AI落地的窗口期布局。

对于做外贸供货的工厂而言,本文指明了数字化转型的方向和新商业机会,干货内容如下:

1. 产品采购和匹配逻辑已经变化,现在海外采购已经转变为AI驱动模式,采购商直接用自然语言提需求,AI会完成趋势发现、产品打磨、供应商筛选全流程,工厂需要尽早规范产品信息,适配AI的匹配规则,才能更容易被精准匹配到,获得更多订单。

2. 工厂推进数字化和电商转型有了低成本工具,Accio作为全球电商AI工作台,现有1000万中小企业在用,打通了国内外多个主流电商渠道,而且Accio的AI token成本比通用Agent低50%,中小工厂也能日常使用,不用投入高额成本就能完成数字化升级。

3. 未来B2B贸易会走向A2A模式,工厂可以提前对接AI工具,适配新的贸易形态,抓住新一轮贸易升级的机会,提升自身竞争力拿到更多海外订单。

对于AI和电商服务商而言,本文透露了行业发展趋势、客户痛点和可行方向,干货内容如下:

1. 行业已经进入新的发展阶段,B2B贸易未来是A2A也就是Agent对Agent的模式,目前已经进入落地爆发期,相关数据显示,Accio的AI token消耗量四个月增长6倍,全球用户一年增长超30倍,说明市场对商用AI的需求非常旺盛,赛道增长空间大。

2. 目前行业的核心客户痛点是,多数通用AI成本过高,中小企业不敢日常使用,同时行业缺乏面向真实电商场景的Agent能力测试标准,很多AI能力强但没法实际跑通生意流程。

3. 可行的解决方案方向可以参考张阔提出的公式:Agent=模型能力×执行框架×业务上下文,同时可以基于开源的CommerceAgentBench测试优化产品,按任务路由匹配不同模型,就能降低AI成本,Accio已经实现token成本比通用Agent低50%,方向可参考。

对于跨境电商平台商而言,本文指明了AI时代平台的发展方向和运营要点,干货内容如下:

1. 当前商家对平台的核心需求已经变化,商家需要能统一管理多平台生意的工具,同时需要能用、用得起的商业AI服务,现有很多AI产品成本过高,无法满足中小商家日常使用的需求。

2. AI时代平台可参考的新做法是,推出一体化AI工作台模式,基于平台真实的业务数据训练优化AI,通过开源基准测试推动整个行业的AI能力升级,同时采用按任务路由匹配不同模型的方法降低AI使用成本,就能做到比通用Agent低一半的token成本,让中小商家用得起。

3. 平台招商可以把AI工具作为核心卖点,目前Accio一年用户增长超30倍,已经拥有1000万中小企业用户,说明AI工具对商家吸引力很强,同时要规避AI成本过高不实用的风险,聚焦实用化、低成本的方向开发服务。

对于产业和AI领域的研究者而言,本文披露了跨境电商AI落地的最新产业动向和理论创新,干货内容如下:

1. 全球跨境B2B产业已经出现新动向,AI落地进入实质性爆发阶段,海外采购商行为已经发生结构性变化,从传统关键词搜索转向自然语言需求描述,Accio的AI token消耗量四个月增长6倍,全球用户一年增长逾30倍,现有1000万中小企业使用,说明商业AI已经从概念走向实际应用。

2. 领域内出现新的理论判断和总结,提出B2B贸易的未来是A2A也就是Agent对Agent做生意的新判断,还给出了可用商业Agent的公式:Agent=模型能力×执行框架×业务上下文,为后续商业AI的开发研究指明了方向。

3. 行业出现标志性新进展,开源了全球首个面向真实电商场景的Agent基准测试CommerceAgentBench,基于千万级用户的真实对话和执行轨迹提炼,解决了行业缺乏真实场景测试标准的问题,为后续研究提供了基础数据支撑。

返回默认

声明:快读内容全程由AI生成,请注意甄别信息。如您发现问题,请发送邮件至 run@ebrun.com 。

我是 品牌商 卖家 工厂 服务商 平台商 研究者 帮我再读一遍。

Quick Summary

This article summarizes key AI e-commerce announcements made by Alibaba.com at its CoCreate 2026 Global Buyer Summit:

1. Alibaba.com CEO Zhang Kuo proposed that the future of B2B trade is A2A (Agent-to-Agent) commerce, a model that has already been launched. Overseas buyers can now describe their procurement needs in natural language, and AI handles the entire process end-to-end: trend research, product selection, supplier screening, contract negotiation and order management. Data shows that traffic to AI-powered multimodal search has grown by over 132%, and global users of Alibaba.com's AI workspace Accio have surged more than 30-fold in one year, with 10 million SMEs currently using the tool.

2. For cross-border practitioners, Accio delivers tangible practical benefits: it has integrated all major domestic and international e-commerce platforms, including Taobao, Tmall, Amazon and TikTok, enabling unified multi-platform business management. Additionally, Accio's AI token cost is 50% lower than that of general-purpose agents, making it affordable for SMEs for daily use.

For cross-border brands, this article outlines new consumer trends and business opportunities in the B2B space:

1. A structural shift has occurred on the procurement side: buyers are moving from keyword-based search to describing needs in natural language, with AI multimodal search traffic growing more than 132% – indicating that AI-powered procurement has already been widely adopted by users. Brands need to adjust their traffic operation and product display strategies to adapt to this new interaction model.

2. Brands can gain multi-channel operation advantages via the new tool: With 10 million SME users, Accio integrates all major domestic and international e-commerce platforms, allowing brands to manage business across multiple platforms in one place and cut operational costs. With AI token costs 50% lower than general-purpose agents, it can be used daily without creating excessive cost pressure.

3. B2B trade will inevitably move toward the A2A model, and brands that make early preparations for AI-driven automated transactions can secure an early lead in this new track and keep pace with the evolving trade landscape.

For cross-border e-commerce sellers, this article reveals new growth opportunities and operational directions:

1. Market demand has shifted: Overseas buyers have changed their procurement habits, moving from traditional keyword search to describing needs in natural language. The 132%+ growth in AI multimodal search traffic confirms that AI-powered procurement is the new traffic trend. Sellers need to adjust their operational logic in time to adapt to new rules and avoid missing out on traffic opportunities.

2. Sellers can get practical tool support: In just over a year since launch, Accio's global user base has grown more than 30-fold to 10 million SME users. It integrates all major domestic and international e-commerce platforms, allowing sellers to manage multi-platform business in one place without switching between multiple backends. Its token cost is also 50% lower than general-purpose agents, making it affordable for small and medium-sized sellers for daily use.

3. B2B trade will shift to the A2A model in the future, and sellers that adopt AI tools early will gain an edge in matching and secure more orders. Sellers should take advantage of the window of opportunity for AI commercialization to get an early start.

For foreign trade-oriented manufacturers, this article clarifies the direction for digital transformation and new business opportunities:

1. The logic of product procurement and matching has fundamentally changed: Procurement is now AI-driven, with buyers submitting needs via natural language, and AI completing the full process of trend identification, product refinement and supplier screening. Manufacturers need to standardize their product information and adapt to AI matching rules as soon as possible to be accurately matched with buyers and win more orders.

2. Manufacturers now have access to a low-cost tool to advance digital and e-commerce transformation: As a global AI workspace for e-commerce, Accio currently serves 10 million SMEs and integrates all major domestic and international e-commerce channels. With AI token costs 50% lower than general-purpose agents, small and medium-sized factories can use it daily and complete digital upgrades without heavy upfront investment.

3. B2B trade will evolve toward the A2A model. Manufacturers can integrate AI tools early to adapt to the new trade landscape, seize the opportunities brought by the new round of trade upgrading, improve their competitiveness and win more overseas orders.

For AI and e-commerce service providers, this article outlines industry trends, core pain points and viable development paths:

1. The industry has entered a new stage of development: The future of B2B trade is the A2A (Agent-to-Agent) model, which is already in a period of explosive commercial adoption. Data shows that Accio's AI token consumption grew 6-fold in four months, and its global user base grew more than 30-fold in one year, indicating extremely strong market demand for commercial AI and huge growth potential in the track.

2. The core pain points facing the industry today are twofold: most general-purpose AI solutions are too expensive for SMEs to use on a daily basis, and the industry lacks agent capability testing standards tailored for real-world e-commerce scenarios – meaning many AI products boast strong technical capabilities but cannot actually complete end-to-end business processes.

3. Viable development directions can draw from the framework proposed by Zhang Kuo: Agent = Model Capability × Execution Framework × Business Context. Providers can also optimize products based on the open-source CommerceAgentBench benchmark, and match different models via task routing to cut AI costs. Accio has already achieved 50% lower token costs than general-purpose agents, proving the viability of this path.

For cross-border e-commerce platforms, this article lays out development directions and operational priorities for the AI era:

1. Merchants' core demands for platforms have changed: They now need tools that enable unified multi-platform business management, as well as accessible, affordable commercial AI services. Most existing AI products are too costly to meet the daily needs of small and medium-sized merchants.

2. For platforms looking to adapt to the AI era, a proven new model is launching an integrated AI workspace, training and optimizing AI based on the platform's real business data, driving industry-wide AI capability improvement via an open-source benchmark, and cutting AI usage costs through task-based model routing. This approach has allowed Accio to achieve 50% lower token costs than general-purpose agents, making it affordable for small and medium-sized merchants.

3. Platforms can position AI tools as a core selling point for merchant recruitment: Accio's 30-fold user growth and 10 million-strong SME user base in just one year proves that AI tools have strong appeal for merchants. Platforms should also avoid the pitfall of overpriced, impractical AI, and focus on developing practical, low-cost services.

For industry and AI researchers, this article discloses the latest industry developments and theoretical innovations for AI adoption in cross-border e-commerce:

1. A major new development has emerged in the global cross-border B2B industry: AI adoption has entered a stage of substantive explosive growth. Overseas buyer behavior has undergone structural change, shifting from traditional keyword search to natural language demand description. Accio's 6-fold growth in AI token consumption in four months, more than 30-fold user growth in a year, and 10 million current SME users confirm that commercial AI has moved from concept to real-world application.

2. New theoretical judgment and frameworks have been put forward: The new proposition that the future of B2B trade is A2A (Agent-to-Agent) commerce, accompanied by a formal formula for practical commercial agents: Agent = Model Capability × Execution Framework × Business Context, which clarifies the direction for future commercial AI research and development.

3. A landmark industry progress is the open sourcing of CommerceAgentBench, the world's first agent benchmark for real e-commerce scenarios. Derived from real user conversations and execution tracks from 10 million+ users, it solves the industry's long-standing lack of real-scenario testing standards and provides fundamental data support for future research.

Disclaimer: The "Quick Summary" content is entirely generated by AI. Please exercise discretion when interpreting the information. For issues or corrections, please email run@ebrun.com .

I am a Brand Seller Factory Service Provider Marketplace Seller Researcher Read it again.

日前,阿里国际站一年一度的海外买家大会——CoCreate 2026,在美国洛杉矶会展中心开幕,当地超15000家中小企业到场。

“B2B的未来是A2A——Agent对Agent做生意。”阿里国际站总裁张阔在主论坛发表演讲,提出了关于全球贸易形态的判断。

这一判断并非纯粹前瞻。据介绍,阿里国际站上的海外采购商行为已经发生结构性变化:他们不再往搜索框里敲关键词,而是直接用自然语言描述需求,随后由AI完成发现趋势、打磨产品、筛查供应商、谈判条款、管理订单流程。

数据显示,在阿里国际站上,海外买家使用AI多模态搜索,流量增长超过132%。而在阿里国际站旗下的全球电商AI工作台Accio上,AI token消耗量在短短4个月里增长了6倍。

过去一年,Accio的全球用户增长了逾30倍,目前已经有1000万中小企业在用Accio做生意。这次到场的15000名参会观众中,有不少就是第一批用Accio这样的Agent做跨境采购的。

在峰会现场,Accio还宣布在GitHub上开源CommerceAgentBench。这是一套面向真实电商场景的Agent基准测试。其数据底座来自Accio上的1000万活跃中小企业用户、160万条真实对话与20万条执行轨迹,最终提炼为107个真实的端到端商业任务。正如张阔所言,CommerceAgentBench测的是一个AI到底能不能真的把生意跑起来。

针对“Agent如何才能真正可用”,张阔还给出了一个公式:Agent=model(模型能力)x harness(执行框架)x context(业务上下文)。这也解释了Accio为什么要开源电商基准测试,其实就是为不同任务匹配更适合的模型,也就是按任务路由不同模型,并用真实商业工作对自研模型做后训练。

基于此,在核心任务案例上的测算显示,Accio的token成本比通用Agent低50%。“对一家小企业来说,一个贵到不敢天天用的AI,等于没用。”张阔表示,“我们的目标不只是把AI做得更强,是要让商业AI真正实用、真正用得起。”

此外,值得一提的是,继在国内打通淘宝、天猫、拼多多、京东、抖店、1688等多个电商平台后,这次Accio又打通了Amazon、ebay、shopify、Temu、TikTok等多个海外电商平台,真正实现“多盘生意,一个AI工作台”。

本文首发于 亿邦动力 官方网站

文章来源:亿邦动力

广告
微信
朋友圈

FAQ回顾

阿里国际站Accio是什么?

Accio是阿里国际站旗下的全球电商AI工作台,过去一年全球用户增长逾30倍,目前已有1000万中小企业使用,已打通淘宝、京东、Amazon等国内外主流电商平台,可实现多盘生意统一管理,token成本比通用Agent低50%。

CommerceAgentBench是什么?

CommerceAgentBench是阿里国际站Accio开源的首个面向真实电商场景的Agent基准测试,数据底座来自1000万活跃中小企业用户、160万条真实对话与20万条执行轨迹,提炼出107个真实端到端商业任务,可测试AI能否实际开展生意。

当前B2B跨境贸易发展有什么新趋势?

当前B2B跨境贸易正朝A2A(Agent对Agent做生意)方向发展,海外采购商已普遍使用自然语言描述需求,由AI完成趋势发现、供应商筛查、谈判、订单管理等全流程工作,阿里国际站AI多模态搜索流量已增长超132%。

商业AI落地需要满足哪些要求?

商业AI落地需要做到真正实用、用得起,Agent能力等于模型能力、执行框架、业务上下文三者的乘积,企业可通过按任务适配模型、用真实商业数据后训练降低使用成本,目前Accio的token成本比通用Agent低50%。

这么好看,分享一下?

朋友圈 分享

APP内打开

赞 +1
+1
微信好友 朋友圈 新浪微博 QQ空间
关闭
收藏成功
发送
/140 0