阿里云张国华谈AI代理与全球贸易的未来 - 第291期

你好，欢迎收听NVIDIA人工智能播客。

Hello, and welcome to the NVIDIA AI podcast.

我是您的主持人诺亚·克拉比茨。

I'm your host, Noah Krabitz.

自1999年以来，阿里巴巴国际站已为来自全球200多个国家和地区的B2B买家和供应商提供服务。

Since 1999, alibaba.com has served business to business ecommerce buyers and suppliers from over 200 countries and regions around the world.

今天，阿里巴巴国际站总裁张阔与我们一同探讨，像阿里巴巴最新推出的Accio这样的AI代理，将在未来十年如何重塑我们庞大的全球贸易生态系统。

Kuo Zhang, president of alibaba.com, is here with us today to talk about how AI agents like Alibaba's recently launched Accio will be reshaping our massive global trade ecosystem over the next decade.

尽管我们在最近的播客节目中已经讨论了大量关于代理的话题，但世界上很少有人能像张阔这样，深刻理解技术如何塑造全球贸易，因此我非常兴奋能邀请他来到我们的播客。

And while we've talked a lot about agents in recent episodes of the podcast, there are few people in the world who have Kuo's perspective on how technology shapes global commerce, which is why I'm so excited to welcome him onto the podcast.

刘先生，非常感谢您抽出时间参加NVIDIA的人工智能播客。

Kuo, thanks so much for taking the time out to join NVIDIA's AI podcast.

谢谢您的邀请。

Thank you for having me.

我知道您是在旅途中抽空加入我们的，因此特别感谢您。

And I know you're joining us while you're traveling, so an extra special thanks.

感谢你协调时间过来。

Appreciate you making it work.

那我们直接进入正题吧。

So let's get right into it.

也许我可以先做个铺垫，或者你来介绍一下。

Maybe we can set the stage or you can set the stage.

请给听众简单介绍一下阿里巴巴是做什么的，特别是alibaba.com，然后你可以谈谈你作为alibaba.com总裁的角色。

Tell the audience a bit about what Alibaba does and alibaba.com in particular, and then you can talk a little about your role as president of alibaba.com.

好的。

Okay.

alibaba.com是阿里巴巴集团的第一个业务。

So alibaba.com is the first business of Alibaba Group.

它由马云和18位创始人于1999年创立。

It is founded in 1999 by Jack Ma and the 18 founders.

最初它只是一个黄页网站，如今已发展成为全球领先的B2B平台。

So it started as a yellow page, and now it's kind of evolving to a leading b to b platform in the world.

它每年连接约五千万名B2B买家和超过二十万家全球供应商。

It connecting around 50,000,000 buyers, b to b buyers in yearly basis, and more than 200,000 suppliers globally.

它目前每年促成的交易额超过六千亿美元。

And it's now enabling more than 60,000,000,000 US dollars transaction in yearly basis.

这就是今天阿里巴巴国际站的情况。

So this is about alibaba.com today.

我实际上从2011年就开始加入阿里巴巴。

And I actually, I joined Alibaba since 2011.

我最初的职位是在淘宝和天猫，主要负责中国国内的B2C业务。

So my first role is in Taobao and Tmall, is kind of in China domestic b to c business.

我从2017年开始加入阿里巴巴国际站，并在五年前成为总裁。

And I joined alibaba.com since 2017 and become the president about five years ago.

所以，这真是一段不平凡的旅程。

So, yeah, this is a quite a journey.

是的。

Yes.

我只能想象，我的意思是，我不知道你是否能在短时间内谈到这一点，但我只能想象，从技术角度，以及你自己的视角来看，买家和供应商方面发生了多大的变化，当然还有阿里巴巴在构建平台方面所做和正在做的事情。

I can only imagine how I mean, maybe I don't know if you can speak to this in in a short amount of time, but I can only imagine how technology and from your perspective has changed so much, you know, on the inside, on the buyer and supplier side and and, of course, what Alibaba has done and is doing building platforms.

因为作为消费者，我在供应链末端的购物方式已经发生了翻天覆地的变化。

Because as a consumer, the way that, you know, I purchase things way down at the end of the chain has changed so much.

所以从你的角度，以及你所做和仍在继续的工作来看，我只能想象这一切发生了多么巨大的转变。

And so I can only imagine from your perspective and the work that you've done and continue to do now, just really how how things have transformed.

没错。

Right.

你说得对。

So what you say is is correct.

我想附和这一点。

I want to echo that.

我们的一个梦想就是让全球贸易像网上购物一样简单。

It's one of our dream is to make global trade as easy as online shopping.

是的。

Yeah.

所以想想我们如何重塑电子商务，让购物变得非常简单。

So think about how we can reshape about the e commerce, so the people buying stuff is really easy.

现在，当买家想要从海外供应商采购或购买商品时，他们仍然会面临许多挑战，包括语言障碍、时差、文化差异和信任问题。

And now when the buyers want to source or want to buy something from a kind of supplier, overseas especially, it's still they will will meet a lot of challenges, including the language barrier, time difference, the time zones, and the culture difference, the trust issues.

那么如何支付、如何结算款项、如何处理物流、如何提供售后服务等等。

So how to kind of pay, how to settle the payment, how to settle the logistics, how to settle the kind of after sales services, so on and so forth.

而且，确实还有很多事情需要依靠技术来完成。

And, yeah, so there's a lot lot of things to be done by technology.

在人工智能出现之前，我们其实已经建立了最初的供需系统、搜索功能、产品展示以及在线沟通，比如直播。

So before AI, actually, we already set up kind of the first, the demand and supply system, the search and the product listings and the online communications, kind of the the live live shows.

然后我们构建了交易系统，包括支付和网络支付系统，以及物流网络。

And then we build up the transaction systems, including the payment and the network payment networks and the logistic networks.

好的。

Okay.

接着建立了完整的信任体系，以支持B2B场景中买家和供应商之间的额外支付。

And then set up the entire trust systems to kind of the b two b extra payment for the buyer and suppliers in in the b two b scenarios.

所以我认为，我们已经建立了一个用于B2B的标准电商平台。

So that's I think we we already set up a kind of a standards e commerce platform for the b two b.

现在我们知道，AI时代即将到来。

And now we we know that AI era is coming.

因此，我们也看到了这里还有很多改进空间。

So we see a lot of improvement space in here as well.

这就是我们推出Axiom的原因。

That's why we in introduced Axiom.

那我们来谈谈Axiom吧。

So let's talk about Axiom.

根据我阅读的资料和在网上看的一些视频，它被描述为一个帮助你做业务的AI代理，听起来很简单，但正如你所暗示的，做B2B业务，尤其是在全球范围内，涉及的内容非常多。

From the materials I read and I looked at some videos online, it's described as an AI agent designed to help you do business, which sounds simple enough, but as you've alluded to, there's a lot that goes into doing business b to b, particularly on a on a global scale.

那么你能告诉我们Axiom具体做什么，以及它如何融入其中吗？

So can you tell us what does Axio do and how does it fit in?

为什么它对阿里巴巴的整体愿景如此重要？

Why is it so important to alibaba.com's overall vision?

当然。

Sure.

所以我可以和你分享我们的愿景，我们为什么这么做，以及一些数据来证明它。

So I can share with you our kind of vision, why we do it, and I can share you some data to prove it.

完美。

Perfect.

Axiom实际上是一个原生AI应用。

So Axial actually is an AI native application.

当我们构建这个应用时，我们基于最先进的模型进行开发。

So when we build this application, we build it on top of the kind of SOTA model.

我们发现，当人们现在使用Axiom时，他们的行为方式与之前使用传统搜索引擎或传统平台时不同。

We see when the people are now using Axial, it's a kind of the behavior is different, like they're using the traditional search engine or traditional platform before.

首先，他们会使用自然语言或长句子来描述他们的需求。

First four, they are using the kind of natural language or long sentences to describe about their request.

以前，他们可能会用类似‘我想买一个便携式储能设备’这样的语言。

So previously, they may use the language like I want to buy a kind of how to say, the the portable energy storage to buy something like that.

像发电机还是电池？

Like a a generator or a battery?

是的。

Yeah.

就像电池一样。

It's like a battery.

电池。

Battery.

好的。

Okay.

好的。

Okay.

现在他们可以用完整的句子来描述这种电池的使用场景。

And now they can describe in a kind of full sentence, like, what is the scenario this battery's usage?

那么，平板电脑长什么样？

So what a tablet look like?

就像一个手提箱，它是便携的，具有什么样的保护措施，尺寸和重量是多少。

Is like a kind of suitcase, is it portable with what kind of protection and the dimensions of the size, the kind of weight.

你可以将它组合起来，然后实际的引擎能够理解你的需求，并将其分解为不同的元素，匹配产品和供应商。

You can put it together, and then the actual engine can understand what you are requiring and kind of break down into different elements and match the products and match the suppliers.

所以这是一种完全不同的用户场景。

So it's a kind of a completely different user scenarios.

我们发现，人们在Axiom系统中输入的句子越来越长，也更加自然。

And we see that people are putting much longer sentences, more natural sentences into the Axiom system.

我认为这是其中之一。

I think that's one.

第二点是，我们现在正在将代理模型引入Axiom系统。

Second is, now we're introducing agent model, actually agent model to the Axiom.

这不再仅仅是执行搜索功能，而是充当一个代理。

This now is not only doing the search functions, but also is acting as an agent.

也就是说，你可以给它一个非常复杂的任务。

So meaning that you can give them a very complex task.

你可以执行任务，并将结果交付给你。

You can execute upon the task and give you and deliver the result to you.

我们发现，在Axiom系统中，Axiom和阿里巴巴.com的用户覆盖率仅为30%。

And what we see is that in Axio, the user, the audience in Axio and in Alibaba dot com, it's only 30% of coverage.

这意味着，很多人实际上正在使用Axiom进行全球贸易的首次在线采购。

Meaning, a lot of people actually is using Axio to do the first online sourcing for the global trade.

它是一种社区模式，降低了人们进入这一领域的门槛。

And it's a community kind of lower the barriers for people to enter in this field.

所以，你刚才描述了这一点，但让我们再深入一点。

So how does just you you described it, but kind of to dig in for a second.

这种体验有什么不同？

How does the experience differ?

或者你能更详细地解释一下，真正实现全球企业间采购需要哪些环节吗？

Or maybe you can unpack a little more of all the different things that have to go into actually buying an item business to business globally.

你之前提到过，比如建立信任网络、技术基础设施、物流、支付以及其他信任因素。

And you kind of alluded to, you know, building the network of trust, the technical infrastructure, the logistics, payment, other trust factors.

你能描述一下传统上需要手动完成、耗时较长的一些事情吗？尤其是对你所说的刚接触这一领域的人来说，现在Axio可以帮他们自动处理这些任务。

Can you describe maybe some of the things that traditionally have been done manually that can take up quite a bit of time particularly as you you said for someone new to doing this, that now Axio can sort of take care of and automate.

当然可以。

Sure.

我可以给你举两个例子。

I can give you two examples.

好的。

Okay.

这两个例子是我团队告诉我的，就在一两周前，他们观察到使用这个系统的人的情况。

So these two examples actually the it's my team tell me that just one week or two weeks ago, how they see the people that are using the system.

其中一个例子是为Polyvian或Bolivian赛事提供服务的供应商。

So one example is this kind of a supplier for Polyvian or Bolivian games.

这是一种在拉丁美洲举办的迷你版奥运会性质的活动。

It's a kind of mini size of Olympic games helping in Latin America.

好的。

Okay.

大约有六个国家参加这些赛事。

About six countries is attending that games.

对。

Right.

当供应商为这些赛事采购物品时，你可以想象他们需要采购各种产品，从金属、齿轮到服装、防护用品等。

And once the suppliers actually is sourcing the items for these games, you can imagine they are sourcing all different products from metals to to gears, to clothes, to protection stuff.

并且需要符合当地的合规规定。

And it's need to kind of apply to the local compliance regulations.

嗯。

Mhmm.

好的。

Okay.

以前，他们需要一支具备专业知识的团队，从成百上千家供应商那里采购，以支持这类赛事。

Previously, they need a kind of team with expertise to sourcing from all kinds of suppliers, maybe hundreds of suppliers to support such kind of game.

所以现在我们看到的是，他们上传一个类似Excel的文件。

So now what do we see is that they upload a file like Excel.

所以要说明他们需要的所有规格，物品数量有数百甚至上千项。

So telling about all the specifications they need, the items are in hundreds or thousands of them.

你可以将这个文件上传到Axiou，Axiou能理解你的需求，知道这个产品将用于拉丁美洲，需要遵守当地的合规要求和指导，并同时执行这些任务。

And you can just upload this file to Axiou, and Axiou can understand what your requirement, understand this is a kind of a is product is going to be used in Latin America, need to follow the the local compliance and guidance, and then it will act simultaneously executed these tasks.

过去，你可能需要数周甚至数月才能完成这份采购清单。

And previously, you may take weeks, even months to finish this sourcing list.

现在使用Assute，可以在几小时或几分钟内完成。

Now using Assute, it can be finished in hours or in minutes.

它还能为你提供所有能够生产该产品的供应商，并给出建议，你可以据此发送询价。

And then it can give you all the kind of suppliers who can make this product and give you a suggestion, then you can use that to send inquiries.

它甚至能更进一步，直接与这些供应商进行沟通。

And it even can can take this step further to communicate with these suppliers.

我认为这就是一个例子，说明智能代理如何协同合作来帮助你。

I think that is one of the examples I can tell you how the agents can collaborate together to help you.

我再给你另一个例子，这是来自专业领域的需求。

The other I can give you one more is, so this is the kind of requirement from expertise.

那么，谁真正拥有这种采购经验，并需要这个代理来协助完成这些复杂的任务呢？

So who actually have this sourcing experience, does need this agent to help them to execute on the complicated task.

我看到的另一个例子是，有些人只有想法。

The other examples I see is that the people just have idea.

比如，我有一个想法，想为注意力缺陷多动障碍（ADHD）儿童设计服装。

For example, one of the ideas that I want to design the clothes for the ADHD children child.

那么，应该使用什么样的材料，如何设计这种服装，才能帮助ADHD儿童呢？

So what a type of kind of materials it is and how to design this kind of child, which is it can help the ADHD children.

对。

Right.

因此，Axiou可以帮助你从市场调研开始，了解现有产品，并一步步给出建议，然后提供供应商和产品推荐，甚至能协助你进行设计原型。

So then the Axio can help you to start from the marketing research to see what the existing product is and give you suggestions step by step, and then give you the suppliers and the product recommendations, and even can help you with the design prototype.

从市场调研到产品设计或重新设计，再到寻找能够生产该产品的供应商，所有这些工作都可以由这个代理完成。

So all this kind of stuff from the marketing research to the product design or product redesign, or the kind of find the suppliers who can make the product, all this stuff can be executed by this agent.

所以这几乎涵盖了整个商业生命周期。

So it's really almost the full business life cycle.

没错。

Exactly.

你简直猜中了我的想法，几乎立刻就说到我想的了，那就是它能执行到什么程度？

And so you sort of read my mind, you almost immediately got to what I was thinking, which was how far down the line can it execute?

你提到过，比如Axio可以提供供应商信息，然后我会发送询价。

You kind of mentioned, you know, providing that Axio could provide the sources and then, you know, I would send inquiries.

但实际上，系统可以直接发送询价，实现代理之间的协作。

But actually, the system could send the inquiries and agent to agent collaboration.

你能再多讲讲这方面吗？

Can you talk a little bit more about that?

无论是它现在能做什么，还是你正在开发的功能，或者未来可能实现的功能，如果你能谈谈的话。

Either what it can do now or sort of what you're you're working on or is possible, if you can speak to that.

好的。

Okay.

首先，我可以告诉你我们是如何设计这个系统的。

So first, I I can tell you the how we design the system.

是的

Yeah.

然后我可以告诉你关于边界的事情。

And then I can tell you about the boundaries.

比如，我觉得你对我们的方向有很多问题。

Like, I think you you have a lot of questions at where we go.

对。

Right.

我们的方向是？

Where we'll Right?

所以系统是这样的。

So the system is looking like this.

首先，你从你的问题开始。

So you first start from your questions.

你发送一个请求，可能是自然语言形式，也可能是多模态的。

You send a request, either it's a kind of in natural language or it's a multi model.

对。

Right.

你发送一份图纸、设计，或者某种文件，比如 PDF 或 Excel 表格。

You send out a drawing, a design, or kind of a file, PDF, Excel list.

它会理解你的需求，然后将其分解为一系列可执行的任务。

It interpret your request and then orchestrate into a kind of set of tasks that you can execute upon.

这可以适用于已经运营、具有复杂采购需求的企业，或者像你提到的拉丁美洲游戏的例子。

That can be at the level of a a business already running that has sophisticated sourcing needs or like the example you mentioned with the games in Latin America.

或者它也可能适用于个人，比如你提到的，有人有设计适应多动症人群的触感服装的想法。

Or it could be something like an individual, as you said, for instance, who has an idea for clothing design, tactile clothing for ADHD wears.

所以，即使一个人对采购一无所知，系统也能帮助他们进行设计。

And so it could be somebody who doesn't know anything about sourcing even as you said, it could help the system can help them design.

因此，无论用户具备何种程度的专业知识都可以使用。

So any level of expertise coming in.

没错。

Exactly.

好的。

Okay.

对于这一部分，实际上需要人类用户来解读。

And for that part, actually, it can involve the human, actually the users interpret.

对。

Right.

所以这些是为你编排或设计好的任务。

So like here are the tasks orchestrated or designed for you.

你想检查什么？

What do you want to check?

如果不需要，系统就会利用所有SATA模型来执行这些任务，为你提供最佳结果。

If not, then it will execute upon these tasks using all the SATA model to give you the best result.

但那是第二步。

But that's the second step.

第三步是重新评估。

And the third step is to reevaluate.

在许多情况下，这种全球采购或全球贸易场景中，决策并不总是定量任务。

So in many of the cases, in this kind of global sourcing or global trading scenarios, so the decisions is not always quantitative tasks.

很多时候，它也是定性任务或决策。

Many of the time it's qualitative tasks or decisions as well.

他们需要评估我们提供的决策或答案是否正确。

They need to evaluate whether the decisions or the answers that we give is the right answer.

平台如阿里巴巴会验证这是否是一个好的答案或好的输出。

It will be proved by the platform like alibaba.com to see if this is a good answer or if this is a good output.

如果不是，我们将重新迭代整个过程，看看如何改进并帮助人们。

If not, then we will iterate the whole process and see how we can kind of evolve and help the people.

每当有决策无法由机器做出时，比如在达成交易时。

And whenever actually, there's a decision cannot be made by, let's say, the machines, like when you make a deal.

那么你能提供哪些具体条件，就需要人类来做出最终决策。

So what precise, what kind of conditions that you can offer, then it will involve the human to make the final decisions.

或者当AI超出其能力边界、缺乏相关知识时，它会返回给人类，以确保他们理解并能够执行。

Or when the kind of the AI actually exceeds its boundary, it does not have this kind of knowledge, it will come back to the humans to make sure that they understand and can execute upon.

这就是我们如何执行整个系统的方式。

So this is how we kind of execute this whole system.

你提到的例子中，为世界某个特定地区的一个活动采购材料。

You mentioned in your example, sourcing materials for an event in a particular part of the world.

有些情况是特定于情境、文化和地点的，即使不同项目表面上看起来相似，也需要处理这些差异。

Some of the just situation specific, cultural specific, location specific things that have to be dealt with for different projects even if they might look the same sort of on the surface.

你如何为全球受众构建系统？

How do you build for a global audience?

阿里巴巴，alibaba.com 已经为全球受众服务了一段时间。

Alibaba, alibaba.com been serving a global audience for some time now.

但当你构建人工智能系统、代理系统时，如何确保这些由AI驱动的解决方案能在这些多样化的市场和用例中有效运行？

But when you're building AI systems, agentic systems, how do you ensure that the AI driven solutions work across these diverse markets and use cases?

是的。

Right.

所以前四个是世界模型和领域特定模型的结合。

So the first four is a kind of a combination of the the world model and the domain specific models.

所谓世界模型，是指当某个国家或地区有特定规范时，这些规范或规则会被模型学习掌握。

The word model meaning that so when there's a kind of specification for a kind of country or region, this specification or the rules will be learned by the model.

因此，当你向该地区发货时，你需要适用哪些法律或规定。

So when you kind of ship product to that area, so what what kind of laws or regulations that you need to apply to.

而与特定领域相关的信息或知识，则由阿里巴巴.com维护。

And the domain specific kind of information or the knowledge is maintained by alibaba.com.

你知道，我们管理了200多家供应商，他们上传了各种证书和不同的产品详情，让我们能够理解。

So you know that we have more than managed 200 different suppliers, and they upload all kind of certificates and all different kind of product details that we can understand.

通过这些信息，我们知道如何将不同产品应用于不同场景和不同国家。

And together, we know that how we can apply different product into different scenarios, into different countries.

此外，这还是一个人机交互系统。

And also, it's kind of a human and the machine interaction system.

在某些情况下，如果超出了AI系统或知识库的边界，我们会介入人工来做出决策或判断。

So in some of the cases that we involve human to make the decision or make the kind of a judgment as well, if it's already exceeds the boundary of the AI system or the knowledge base.

第三点，也是不可忽视的一点是，阿里巴巴.com已经建立了一套系统，确保我们遵守所有法规和规则。

And the third, not the least, is that we already actually, alibaba.com already managed a system, which we comply with all the regulations, so all the rules.

而且，我们已经成功处理了超过600亿美元的交易。

And also, we already managed to process more than 60,000,000,000 US dollar kind of transactions.

所以我们知道如何建立一种系统性方法来保护。

So we know that how to build a a systematic approach to protect.

因此，这三层体系共同作用，能够为全球贸易带来更好的成果。

So these three layers actually together that we can deliver a better result for the global trading.

当涉及到更细致、更偏向定性而非定量的信息、洞察力或文化差异时，比如人们可能本能地知道并据此行动，却难以或从未想过用语言表达出来的东西。

When it gets into more kind of nuanced or or maybe sort of qualitative as opposed to quantitative information, insights, cultural nuances, you know, just things that a person might sort of know and act on intuitively but maybe have a hard time or just never think to express in words.

你们在训练系统时也采用同样的方式吗？

Do you approach training the systems in the same way?

这些类型的知识和能力是否更依赖于在使用中学习？

Are those types of knowledge and capabilities kind of more dependent on learning by use?

你们如何将那些对促成交易至关重要的本地化细微信息融入系统中？

How do you infuse the systems with, you know, the sort of local nuanced information that can be really important to closing deals?

好的。

Okay.

这是个非常好的问题，我认为我们也在持续在这方面努力。

So this is a very good question, and I think we are keeping working on that as well.

所以我们现在所做的，可以说分为三层。

So what we're doing now is, can say that is three layers.

第一层当然是数据集。

The first layer, of course, is the dataset.

我们知道我们拥有超过六亿种产品、供应商、交易和买家，通过这些数据，我们能够了解供需状况，并从中提炼出领域知识。

So knowing that we have more than two sixty million products and suppliers and transactions and the buyers that we, through this data, that we understand what the demand and supply look like, and that we can kind of abstract the domain knowledge from that as well.

我认为这就是第一层。

So I think this is the first layer.

第二部分是行业经验。

The second part is about industry know how.

这一部分不仅依赖于模型，还依赖于行业专业知识。

So, that part actually is based on the, not only the model, but also the industry expertise.

我可以给你举个例子。

So I can give you example.

当我们说为特定的适配场景设计产品时，如何评估这个设计是否优秀呢？

So when we say we design a product for a specific fit scenario, so how we evaluate that design is a good design.

这个问题就像是：我们如何判断一个答案是否优秀？

This is like the the question is, how we evaluate the answer is a good answer.

因此，需要依靠专家协助来评估所有这些答案，并持续改进系统。

So that need to be kind of rely on the expert assistance to evaluate all these answers and to kind of keep improving the system.

对的。

Right.

第三点是关于平台本身。

And the third is about the platform itself.

当我们根据需求得出结果后，会将这个结果放到我们的平台，比如阿里巴巴国际站上。

So when we outcome kind of a result based on the requirement, and then we will put that result into our platform like alibaba.com.

然后我们会观察转化率，看看买家和卖家如何与这个输出互动。

And then we will see the conversion rate and we'll see how the kind of the buyers and the sellers are iterate with this output.

如果效果不好，我们就会迭代系统、模型和数据集。

So if this is not good, and then we will iterate the system, the models, the dataset as well.

这就是我们现在解决这些问题的方法。

So that is how we can now solve these problems.

你可以想象，我们正在利用我们的数据和行业经验，但需要根据系统的反馈实时迭代系统。

So you can imagine that we are leveraging our data, our kind of industry know hows, but we need to iterate the system in real time based on the system's feedback.

我正在与张阔交谈。

I'm speaking with Kuo Zhang.

张阔是阿里巴巴.com的总裁，我们一直在讨论Accio。

Kuo is president of alibaba.com, and we've been talking about Accio.

它们是相对较新的。

They're relatively new.

Accio是什么时候推出的，张阔？

When did Accio launch, Kuo?

第一个版本去年上线，但智能代理版本是上个月刚刚推出的。

The first version launched last year, but the agentic model, the agent version last launched just last month.

上个月。

Last month.

好的。

Okay.

谢谢。

Thank you.

我本来不想说得不准确，既然有这么完美的知识来源在这里，我当然要问清楚。

I I I didn't wanna be imprecise when I have the perfect source of knowledge right here to tell us.

所以上个月，Accio代理发布了，它不仅基于Accio本身，还像你所说的那样，依托了阿里巴巴多年服务全球商业社区所积累的庞大数据库和知识库。

So last month, Accio agent came out building on not only Accio itself, but as you've been saying Alibaba's incredible database and and just knowledge repositories of years of serving the global business community.

我想问问用户端的情况。

I wanna ask about the user end of things.

你之前提到过一些用户行为的对比，比如搜索和使用AI驱动界面之间的差异。

You mentioned before some of the user behavior kind of comparing between search and using the AI powered interfaces and that kind of thing.

但当你谈到那些依赖你们平台开展业务的用户时，正如你所说，他们的业务涉及大量复杂的环节，用户对Accio以及向更智能化的自动化系统过渡的反应如何？自动化一直存在，但在AI时代，似乎用户需要建立一种新的、甚至更高层次的信任。

But when you're talking about users whose businesses rely on your platform for what they're doing and there's a lot of, you know, sophisticated moving parts as you talked about, how have the users responded to, you know, Accio in particular and moving to more, you know AI automated systems, automation's obviously been a thing but now in the AI age, it seems like there's an increased or even new layer of trust that would have to be built with users.

那么，你们是如何建立这种信任的？从技术层面来看，你们采取了哪些保障措施和透明度机制？

So how do you build that trust And on a technical you know, from a technical side of that, what guardrails, transparency measures?

你之前稍微提到了安全性，但在面对代理系统时，你们是如何将这些安全机制融入其中的？

You've talked a little bit about safety, but when you're looking at an agentic system, how do you go about building those guardrails into it?

正如我所说，首先，我认为这始终是人与机器的交互系统。

So as I mentioned, so first of all, I think it's it's always a human and the machine interaction systems.

是的。

Mhmm.

每当我们的AI系统或大型语言模型做出的决策超出其知识边界时，我们都会引入人类来参与决策，比如在达成交易、谈判价格或条件时。

So whenever that we think the decision that we made by the AI system or by the the big larger language model exceeds the boundary or without the knowledge that we have, we always involve humans to make decisions, like when you make a deal, when you're negotiating a price or conditions.

我们始终利用我们的平台来反复优化模型，看看它是否能带来更好的结果。

And we're always using our platform kind of to reiterate the model to see whether it gets a better result.

比如更高的转化率等等。

There's a better kind of conversion rate, so on and so forth.

这基本上就是我们每天在做的事情。

So this is basically we are doing in in daily basis.

你能谈谈中小企业对世界意味着什么，以及它们如何启发了你的工作吗？

Can you talk a little bit about what small and medium sized enterprises mean to the world have meant to and kind of inspired your own work?

然后从你们的业务和这些AI系统的角度来谈谈，正如你所说，这些系统不仅极大地加速了流程——把原本需要一周才能完成的过程缩短到几小时甚至几分钟，而且还降低了准入门槛，让那些单打独斗的创业者或只有想法但缺乏知识、资源和技术能力的团队，现在也能接触到全球市场。

And then kind of talk about that or talk about it through the lens of at your trade and these AI systems that, you know, not only speed up the process incredibly as you were talking about, you know, taking this week's long process, boiling it down to hours, minutes in some cases, but also as you said, make it available, lower that barrier of entry so that, you know, the solo entrepreneur or the team that has an idea but maybe not the knowledge and the resources and the technical skills can now access this global market.

在将这些工具交到中小企业手中时，最让你兴奋的是什么？

What what excites you most about doing this this whole thing and putting these tools into the hands of smaller businesses?

当然。

Sure.

所以我们明天将在拉斯维加斯举办一场图书展。

So you know, we hold a book crate just tomorrow in Vegas.

所以整个团队正在为这件事做准备。

So we are well, the whole team is preparing for that.

好的。

Okay.

对于未来收听或回看的听众来说，时间是九月初。

Early September for listeners listening down the line or talking.

没错。

That's right.

在CoCreatE活动期间，我们举办了一场CoCreatE路演。

And during CoCreatE event, we have a CoCreatE pitch.

好的。

Okay.

今年，我想说，我们收到了超过两万五千份申请，太惊人了。

So in this year, actually, we I wanna say this, more than 25,000 applications Amazing.

仅在三十天内，就收到了这么多CoCreatE路演的申请。

For the co created pitch just in thirty days.

在这些申请中，我认为超过40%的申请者是独立创业者。

And among these applications, I think more than 40% of them mark them as a solo entrepreneur.

好的。

Okay.

这意味着很多人其实都有自己的想法，想要基于全球供应链打造产品和流程，但他们必须独自完成一切——从产品设计到处理客户投诉，再到执行所有物流和融资系统。

So it's meaning that a lot of people actually have their ideas, well, to build their product, build their process based on the global supply chain, but they need to do everything by themselves, from product design to the handle the customer complaints, to execute upon all those logistics, financing systems.

我认为，智能代理模型能从不同角度帮助这些独立创业者。

I think that agentic model can help you help this our solar entrepreneurs, at least in different prospect.

所以我们看到，Axio使用的第一种场景是寻找供应商。

So we see the number one that scenarios are using by Axio is to find suppliers.

就像是，谁能够制造这个？

It's just like, who can make this?

我有个想法。

I have idea.

谁能够制造这个？

Who can make this?

让我们找到他们的商业伙伴。

Let's find their business partner.

他们使用SEO的第一种场景是进行产品设计，或跟随市场上的热销产品，或对市场上的热销产品进行重新设计。

The number one scenario that they use SEO is to have product design or following the winning products in the market or redesign a winning product on the on the market.

所以这都是产品重新设计的部分。

So this is all the product redesign part.

第三点是关于如何找到产品。

And third is about how to find the products.

我认为这些是Axial的三大主要场景。

I think these are the three major scenarios in Axial.

这与其它类型的平台完全不同。

It's completely different from the other kind of platforms.

其它平台主要只是寻找产品、买卖之类的。

The other platforms, majorly, they're just looking for a product, buy and sell, something like that.

但Axial实际上能为他们提供更多的帮助。

But the Axial actually can help them much more.

你知道，当我们1999年创立这个业务时，马云对阿里巴巴集团的使命是让天下没有难做的生意。

And you know, alibaba.com, when we set up this business back in 1999, the Jack Ma's mission for Alibaba Group is to make it easy to do business anywhere.

所以我认为，今天我们为独立创业者所做的，正是在延伸这一使命。

So I think what do we do today with for solo entrepreneurs extend this mission.

有道理。

Makes sense.

在打造Axio、构想并实现一个用于全球贸易的AI系统的过程中，有什么事情让你感到意外吗？

What's been something that surprised you in building Axio and and trying to envision and then bring to life an AI system for global trade?

我认为第一部分是关于技术的。

I think first part is about the technology.

我们今天为全球贸易解决的问题，与B2C电子商务世界完全不同。

So the problem we solve today for global trading is not like the kind of a B2C e commerce world.

在B2C电子商务世界中，当你购买某样东西时，价格通常在几千元人民币，或几美元到几百美元之间。

So in B2C e commerce world, when you buy something, so the price probably within a couple of thousand, a couple of dollars to a couple of hundreds US dollars.

但当涉及到B2B采购，尤其是在全球贸易场景中，问题的成本会变得高得多。

But when it's come to a kind of a b to b sourcing, especially in the global trading scenarios, the questions to become is becoming much more expensive.

而且很多时候，这并不是一个量化的任务，而是需要做出的决策具有高度的定性特征，非常不容易。

And many of the times, it's not quantitative task, it's a kind of qualitative task that decisions that you need to make is not easy.

所以我认为，从技术挑战的角度来看，就是这样。

So I think that is for the kind of technical challenge perspective.

第二部分是关于商业模式的挑战。

The second part is about the business model challenge.

当你在构建一种AI搜索系统时，意味着你不是让用户输入关键词，然后给他们数百万个结果让他们逐个点击。

So you know when you're building kind of AI search, meaning that you are not letting your users key in the keywords, giving them kind of millions of results and letting them to click through.

实际上，你需要理解并整合他们的需求，提供少量结果，并给他们更好的选择。

Actually, you are understanding, integrating their kind of requirements, come with few results, and give them kind of the better choice.

这也可能影响商业模式。

That may impact on the business model as well.

这就像是广告商业模式，需要不断演进。

So it's like a like the advertisement business model that it need to evolve.

对。

Right.

我们知道，这种模式能带来更多的客户价值，进而带来更大的商业价值。

So we know that this model can bring more customer value and then it will bring more business value.

但你仍然需要调整商业模式，以适应这项新技术。

But still, the kind of the business model you need to evolve to kind of match with this new technology.

如果你要向一位高管——无论他们产品的具体类型是什么，但我们可以假设是在电子商务领域——讲述如何构建一个能够大规模部署、规模接近你们日常处理水平的AI产品，

If you were speaking to an executive who's whatever the product may be, but we can say within e commerce, looking to build an AI product deploy at massive scale, scale approaching what what you deal with on a daily basis.

在开始之前，你有什么建议或想法可以分享给他们吗？

Do have a piece of advice or or a couple things that come to mind that you would give them before starting out?

对。

Right.

首先，我认为最重要的是你要解决的问题。

The first, I think the most important one is about the questions you're going to solve.

所有那些关于技术的花哨术语，最终还是要看你的问题是不是一个真实的问题，或者你的问题是不是足够大，我认为这是第一点。

So this that all kind of kind of fancy terms about the technology, is still whether your question is a real question or your question is a big enough question, I think that is the first one.

其次，我可以分享一下我们在过去两三年里积累的一些最佳实践。

And second is that I can share some of the best practice that we experienced for the last two or three years.

因此，我们在实践中采用了三层方法来应用人工智能。

So we have kind of three layers of approach to kind of practice AI.

第一层是原生人工智能应用，也就是Axio，今天我们已经讨论了很多。

The first layer is about AI native applications, which is Axio, we talk a lot today.

在这些原生人工智能应用中，你可以以极快的速度尝试任何你需要的东西，并且能迅速迭代产品。

And in that kind of AI native applications, you can try anything that you need with a very quick speed, and you can reiterate this product very quickly.

第二部分是人工智能与阿里巴巴.com的结合。

And second part is about AI plus alibaba.com.

正如我之前向你们提到的，这是阿里巴巴集团历史长达二十六年的第一个业务。

So which is, as I mentioned to you, which is the first business of Alibaba Group with years of twenty six years history.

我们需要将AI模型或AI价值融入阿里巴巴.com，以实现更大规模的推广。

And we need to add AI model or the AI value to alibaba.com, which can expand in a larger scale.

更多的买家和供应商都能从中受益于这一AI模型。

You can get more people to benefit from this AI model, both the buyers and the suppliers.

这是第二层。

This is the second layer.

第三层是关于AI洞察。

The third layer is about AI insight.

在阿里巴巴.com的各个部门中，每个岗位都有一个AI关键绩效指标。

So within alibaba.com's opposition, every role has an AI KPI.

从用户增长、产品设计，到销售团队、技术团队，等等，阿里巴巴.com组织中的每个岗位都为自己设定了某种AI关键绩效指标。

From the user growth, to the product design, to the sales team, to the technology team, and you can mention, so every role in alibaba.com organization, they have a kind of AI KPI for themselves.

因此，每个人都有一种紧迫感，希望通过AI技术来提升协作效率或改进工作。

So everybody has a sense of urgency to improve either co palette or improve by the AI technology.

KPI是衡量AI的使用情况、生产力，还是AI工具本身的有效性？

The KPI is measuring use of AI or productivity or effectiveness of the AI tool itself?

我认为不同团队或不同岗位

I I think different team or different roles

好的。

Okay.

有不同的KPI。

Have different KPIs.

对。

Right.

比如销售团队，关注的是小机器人效率。

Like in sales team, small bot efficiency.

但在技术团队，更关注的是吞吐量。

But like in technology team, it's more about kind of a throughput.

也就是你能交付多少功能。

So how many features that you can deliver.

在产品团队和用户增长团队中，关键是他们如何利用人工智能重新设计模型和日常工作的流程。

In product team and in kind of in the user growth team, it's like how they can leverage AI to redesign the model, redesign the kind of daily basis work.

我认为整个团队都能从人工智能中获益良多。

I think the whole team can benefit from AI a lot.

这不仅仅关乎某个单一团队或某个人。

It's not only a single team or kind of single person.

对。

Right.

对。

Right.

当然。

Absolutely.

所以，在我们结束之前，如果可以请你展望一下五年、十年甚至更长远的未来，人工智能将如何从根本上改变全球商业的运作方式？

So if I can ask you as we as we wrap up here, to look ahead five years, ten years, some more in that time frame if that works, how is AI going to change just on a fundamental level the way that we do business around the world, global business?

哪些方面会发生变化？特别是，有没有什么你觉得人们可能会感到惊讶的？

What's gonna change and in particular if there's something you think people might find surprising?

我喜欢以这样一个引人深思的问题来结束。

Always like to end on a provocative note like that.

好的。

Okay.

那么，我们该如何定义人工智能的成功呢？

So it means how do we define the success for AI?

当我们谈论API之类的东西时。

When we talk about the API, something like that.

我认为，衡量AI成功的标准在于我们是否能为当前的GDP至少带来10%的增长。

So I think the success of defining AI is whether or not we can add at least 10% of growth on the current GDP.

比如在全球贸易领域。

For example, for global trading.

如今全球贸易的规模超过三万亿美元。

Global trading today is more than 30,000,000,000,000 US dollars business.

如果我们能为这笔业务再增加10%的附加值，就意味着整个系统将新增三万亿美元的价值。

So if we can add 10% more to this business, it's going to be 3,000,000,000,000 US dollars kind of add on value to the whole system.

我们相信，在AI的帮助下，越来越多的人能够预见并融入这种全球供应链，参与全球竞争。

And we believe that with the help of AI, more and more people can anticipate, can embrace this kind of global supply chain and can compete globally.

这将极大地提升整体价值。

That will dramatically kind of increase the value.

正如我们一开始所说，让每个人在任何地方都能轻松做生意。

And as we said in the beginning, to make it easy to do business anywhere for everybody.

很棒。

Cool.

对于想了解更多关于alibaba.com的人，显然可以直接访问网站，但除此之外，关于今天我们讨论的其他方面，比如社交媒体、研究博客或其他资源，你们还有哪些渠道？

For people who would like to know more about alibaba.com, obviously, the website right there, but more about any aspects of what we talked about beyond the website, social media, perhaps there's a research blog, other assets.

听众们应该去哪里了解你和你的同事、团队在阿里巴巴所做的工作？

Where where should listeners go to learn more about the work that you and your colleagues and team are doing at Alibaba?

你可以访问xu.com或alibaba.com，这是首选的入口。

You know, visit the website xu.com or alibaba.com, I think, is the first go to place.

此外，我们还有一个名为'B2B突破'的播客，在美国推出。

And also, we have a kind of a podcast we call the b two b breakthrough in US.

太棒了。

Oh, fantastic.

是的。

Yeah.

我们有很多客户案例和你可以学习的最佳实践。

We have a lot of customer use cases, a lot of kind of best practices that you can learn.

那里有很多内容可以发现。

There's a lot to find there.

很好。

Great.

名字再告诉我一遍，抱歉，是B2B突破吗？

And the name again, sorry, B2B Breakthrough?

B2B突破播客。

B2B Breakthrough Podcast.

完美。

Perfect.