Kuo Zhang, president of Alibaba.com, thinks B2B commerce is about to become A2A — agent-to-agent.
In this AI Transformations conversation, he explains why buyers and sellers will both deploy agents, and what that shift means for global trade. Mike Kaput talks with Kuo about Accio, Alibaba.com's agent platform, the Commerce Agent Bench benchmark, and how a 60-step sourcing process is collapsing from months into hours.
Listen or watch below—and see below for show notes and the transcript.
00:00:00 — Intro
00:7:30 — Why Accio
00:8:49 — Inside Alibaba.com's AI transformation with Accio Work
00:15:12 — The pain points Accio was built to solve
00:19:32 — Real examples of entrepreneurs using Accio Work
00:24:15 — The technology powering the transformation
00:28:50 — How AI agents will reshape commerce at large
00:32:26 — What had to change inside the company
00:37:26 — The hardest parts of shipping agent-based products
00:40:47 — How Alibaba.com measures agent performance
This week’s episode is also brought to you by MAICON, our 6th annual Marketing AI Conference, happening in Cleveland, Oct. 13-15. The code POD100 saves $100 on all pass types.
For more information on MAICON and to register for this year’s conference, visit www.MAICON.ai.
Disclaimer: This transcription was written by AI, thanks to Descript, and has not been edited for content.
[00:00:00] Kuo Zhang: And there are going to be analyst back and forth conversations, different time zones, different cultures and languages in there that AI can help a lot. So we believe that the B2B will go to a to A2A, Agent to Agent, that will be the future.
[00:00:17] Mike Kaput: Welcome to AI Transformations, a special series from the Artificial Intelligence Show.
[00:00:23] I'm Mike Kaput, Chief Content Officer at SmarterX and Marketing AI Institute, and I'll be your host. Every business's AI journey looks different. In each of these episodes, I sit down with leaders who have lived through real AI transformation, including the actual stories, the pain points that push them to act.
[00:00:42] The moment things started to click and the results they can point to today. So join us as we accelerate AI Literacy for All as part of our AI transformations series. Welcome everyone to episode 244 [00:01:00] of the Artificial Intelligence Show. I am Mike Kaput, co-host of the Artificial Intelligence Show and Chief Content Officer here at SmarterX.
[00:01:08] Now, today we are continuing a special series. We've been running called AI Transformations. This is where we spotlight how real companies are driving real change using ai. Now, we actually launched AI transformations with this initial six episode run. Where we had some select stories we knew we wanted to tell, but there are far more than six transformation stories worth telling.
[00:01:36] So we are continuing the series and bringing you more of these conversations directly as part of the artificial intelligence show. Now the premise here remains the same. AI transformation can feel abstract. Plenty of companies talk about it, far fewer show what it actually looks like inside real organizations with real teams, real constraints, and real [00:02:00] business outcomes.
[00:02:00] So in these episodes, we are talking directly to the leaders. Doing the work, we're going to explore their particular AI transformations, what AI work looks like inside the company, what got difficult along the way when it comes to AI adoption and innovation, and what kind of results AI is starting to unlock.
[00:02:19] So today the leader we are talking to is Kuo Zhang, who is president of Alibaba.com Kuo is here to talk about how alibaba.com is using AI and agents to rethink how entrepreneurs and small businesses participate in global commerce. So before we get into today's conversation with Kuo, the. This episode is brought to you by MAICON, our marketing AI conference for marketing and business leaders happening October 13th to the 15th, right here in our home base of Cleveland, Ohio, we are expecting thousands of AI forward leaders here we are less than two weeks away from MAICON.
[00:02:59] If you are [00:03:00] listening to this when it drops. So if you have been thinking about joining us, this is the time to make it happen. We are gonna bring together so many people for three days of keynotes sessions, workshops, and conversations about how organizations are adopting, operationalizing and scaling ai.
[00:03:18] So you'll hear real world case studies, practical frameworks, and direct lessons from the people doing the work inside companies today. If you are a podcast listener, use the code POD100 at checkout for your MAICON ticket to save a hundred dollars on your pass. Now, what's really cool is that using this code will also reserve your place at our private lunch for podcast listeners where Paul Roetzer,
[00:03:43] my other co-host, our CEO and founder and I will spend an hour answering your questions about ai, your company, your career, whatever you're trying to navigate in the wide, wonderful world of ai. So seats are limited. The offer closes when the [00:04:00] room is full. So use that code as soon as possible.
[00:04:02] Go to MAICON.AI MAICON.AI to register, and we hope to see you October 13th to the 15th here in Cleveland. So today I'm joined by Kuo Zong, president of alibaba.com. Now, when most people hear the name Alibaba, they probably think about the much larger Alibaba group. alibaba.com is a specific part of that organization, and it is where the entire.
[00:04:30] Company began. alibaba.com launched in 1999 as Alibaba Group's First Business. Today, it is a global business to business commerce platform that helps entrepreneurs and small and mid-size businesses find suppliers, source products, conduct transactions, and manage different parts of international trade.
[00:04:50] For the fiscal year, 2026, alibaba.com reported that more than 47 million. Buyers from over 190 countries had used the [00:05:00] platform to source business opportunities or complete transactions. It also had more than 240,000 paying members. Kuo has served as president of alibaba.com since 2017. He began his career as a computer scientist at IBM and Microsoft before joining Alibaba in 2011, where he worked on merchant platforms and tools for tbo.
[00:05:21] Now since taking over alibaba.com, he has helped lead its evolution from what was primarily a directory of business listings into a platform that supports transactions, payments, logistics, marketing, and other parts of global commerce. Now he is leading what alibaba.com sees as its next major transformation, the move towards AI powered agent to agent commerce.
[00:05:45] That journey, which we'll talk about today, began in November, 2024 with. Accio, an AI powered search and sourcing engine built for business buyers. It has since evolved into what you can think of as a team of [00:06:00] specialized AI agents for entrepreneurs and small businesses. So rather than answering simply questions or recommending next steps, these agents can break a business goal into specific tasks and help execute them.
[00:06:14] Things like market research and product sourcing to supplier negotiation, store operations, marketing. Logistics and compliance. Now more than 10 million small businesses now use Accio work to run their operations collectively consuming 1 trillion AI tokens over a recent three month period. Now, those are significant numbers, but the bigger story here is what they could represent.
[00:06:39] alibaba.com is exploring now whether it can evolve from a marketplace that people navigate. Into an operating platform where agents increasingly help conduct commerce on their behalf and eventually interact with other agents representing buyers, sellers, and service providers. So we're going to explore today how this [00:07:00] transformation began, what Accio work actually does, how customers are using it, and what had to change.
[00:07:06] Inside alibaba.com. Not to mention, we will chat about what has been difficult along the way and what agent to Agent Commerce could mean for the future of global business Kuo. Welcome to the show. Thank you so much for being here with us.
[00:07:18] Kuo Zhang: Thank you for helping me. You did a lot of research.
[00:07:22] Mike Kaput: We tried, we tried.
[00:07:24] you know, it gets a lot easier these days when agents can help you with the research. I think we will talk about that, I'm sure.
[00:07:30] Mike Kaput: So that's true. Before, before we get into this broader story, this broader transformation, I have to ask about the name, of the product. So Accio, A-C-C-I-O is Latin for I Summon, and Alibaba has actually said in interviews outright, the name is a reference.
[00:07:50] To the summoning charm in Harry Potter. I'm just curious, why was this the right name for what you're building here?
[00:07:55] Kuo Zhang: Yeah, so Accio, what you say is right is [00:08:00] Latin for I summon and also is the kind of the summon charm in Harry Potter. It's just you, kind of say name and you can bring to you. So we use this name because in the beginning when we, when we, design Accio.
[00:08:15] It is like you have a word or you have some thoughts and we can use AI and global supply chain help you to make your dream come true. It's some kind of Accio. And also in Accio there's AI inside this, this term. And all the young guys in a audible.com that kind of we get a vote and all the young guys love, love this, this name.
[00:08:44] So we just choose this as the new product name.
[00:08:48] Mike Kaput: I love that. That's so cool.
[00:08:49] Mike Kaput: So let's start off at a high level. Maybe walk me through. The AI transformation that alibaba.com is pursuing through Accio work. And maybe talk to me a bit about how this is changing the way entrepreneurs and small businesses participate in global commerce.
[00:09:08] Kuo Zhang: Hmm, sure. So we, we take step by step. So Accio first launched I think two or three years ago, and at that time, AI is not as today. We starting from the AI search, you can think about this, like a Google AI mode or AI overview. So when we feel that when the people using alibaba.com, they are using, use the search a kind of, different way than before they are, they are tend to use more and more, multimodal materials like.
[00:09:45] Images like documentations and they describe their needs in a kind of a sentence. Not only a few keywords or queries and leveraging ai, actually we are, we have more, [00:10:00] power to understand the intention behind the scene of the users, buyers on audible com. I think that's the first step. And when we kind of launch the product, we are seeing people actually using this online agent, the online sourcing agent, to do many other things other than search, for example, they are using this agent to develop their ideas.
[00:10:28] They have some ideas. They want to do a market research. They want to see if this product exists. Are there kind of shortcomings of the existing product and what are the opportunities? And then they will transform this idea into a design pack. So previously, when the buyers talk with suppliers back and forth, many of times that they cannot describe what they are trying to build and leverage ai, actually they can generate a, a very sufficient and [00:11:00] professional design pack for the suppliers.
[00:11:02] When the supplier received the pack generated by Accio, they said they never see a buyer actually is so professional. And then it reduced a lot of back and forth discussion and dispute as well. So that is the beginning of the Accio. So we start from the AI search to kind of a sourcing agent, and they're using that not only for inquiries, but also for market research for.
[00:11:32] customer or suppliers waiting for the product design and when the AI, technology or power kind of moving on, we are using AI to do the negotiations part as well. So the AI become a real agent, represents the benefit of the interest of the buyers, and you can imagine the seller side has his own agent as well, helping them to do the similar stuff.[00:12:00]
[00:12:00] And, for this quarter, we can expand the scope of this agent Further, we found that is not only used for sourcing, it can be used as a e-commerce workspace. Remember, I actually, the first job of mine is in top on Timor providing all kind of tools for sellers and we, we see that the AI for e-commerce actually has a lot of similarity of the sourcing.
[00:12:30] So you need to manage many stores like you have Shopify, you have Amazon, you have eBay, et. And many of the sellers that we met actually, they manage. stores across different platforms, but they want to manage the business as one, not in different spreadsheet, understand the p and ls, and also when they update a listing on.
[00:12:56] Shopify on Amazon, on TikTok shop or on social [00:13:00] medias. They want to do it once for all, not kind of keep updating different product to different platforms using different images, videos, applying different rules. And this can be taken care by the agent as well. And the last but not list is that Accio actually is, aggregate all different agents and you can put them in a group.
[00:13:24] Helping you to do some kind of deep research from understanding the trend to understanding your competitors opportunities, and also helping you to dig actually, who can make this and what's the p and l look like? What's the margin look like? And this together for SMBs, like in Los Los Angeles, you have some ideas you want to source.
[00:13:49] You want to list your product in all different platforms and you want to manage your daily business. So this Accio work actually can help you to manage [00:14:00] your end-to-end work, not only on alibaba.com, but also on all your kind of storefronts as well. So this is how this agent evolves.
[00:14:09] Mike Kaput: Wow. So it sounds like Accio started out as essentially product search using AI for better search on alibaba.com.
[00:14:17] But because of what you are seeing. Customers, users actually using the AI for, which wasn't just search, it was ideas. It was advice that it expanded from there to actually become agents that could help with many different parts of the sourcing and e-commerce process when building a business. And it sounds like then that evolved into Accio work being a platform to essentially help you end to end, run and build an e-commerce business.
[00:14:45] Do I have that roughly correct.
[00:14:47] Kuo Zhang: Yes. And since we launch Accio Work, this April, so the tokens that used by the users actually is, sixfold [00:15:00] for the past four month. So that means that they are using these tools much deeper and the complex than.
[00:15:08] Mike Kaput: Yeah, that's an insane amount of tokens.
[00:15:12] Mike Kaput: So before you built Accio and Accio work, I'm curious what was, where was the pain point?
[00:15:19] What was so difficult for entrepreneurs and small businesses? I mean, you kind of mentioned a little bit the agents are advocating or helping, you know. Helping the buyers and the sellers that they're helping communicate between them. I'm curious if you could drill a little more into why were agents, AI agents, the right answer here for the problems that entrepreneurs and sellers and small businesses were experiencing.
[00:15:45] Kuo Zhang: Sure. So before we talk about B2B, we can do a kind of comparation with, B2C. This is more easy for the audience to understand when you buy something. So there are a lot of, [00:16:00] I should say, the experiment or explorations for the B2Cs agent as well, but majority of them is not that successful yet.
[00:16:11] Because of the scenarios or the challenges. So when you a consumer buy some stuff, the first, the product is not that expensive. So decision is not expensive. The second is, the decision. How do you say the dimensions that you need to make that decision? Decision is not complicated. It's in general, including the image title, price, delivery time.
[00:16:37] Probably this is it. And also many of time when you buy something you're killing, time consuming actually is a kill time. Nobody's using agent to kill time. You don't send agent to cinema to see a movie for you, okay? You want to kill time yourself, but in the B2B part, when we do the completion is completely different.[00:17:00]
[00:17:00] So the first, the decision is expensive. So on alibaba.com every order. Kind of, in general is about 3000 US dollars.
[00:17:11] Mike Kaput: Mm.
[00:17:11] Kuo Zhang: So the decision is expensive. And also the demand, the decision, the process of decision is complex. So you need to compare, at least you can think about there's a table, this, at least 20 columns, starting from the name of the product price.
[00:17:35] Then go to the MOQ, then go to the certificate, and then go to the kind of, delivery time, payment terms, inco terms. And a lot of things is other than kind of you buy something for your own use, you want to buy under, re, re resell, or you want to buy because it's part of your product. So it's evolved a lot of, [00:18:00] requirements that are far more complicated than the consumer side.
[00:18:06] And also it's a work, the B2B for both side. They are in a working kind of environment, so they procurement because they need to kind of do a business and suppliers the same. So both part want to use agent to save time, right? For human beings, we save time so that we can kill time. So agent is, designed for same time part.
[00:18:32] So for the B2B, we think is first, is solve a lot of problems, can help you to save a lot of time. And also there are a lot of, complicated decisions, which requires back and forth repetitive work that can be taken care of by the agent. So taking a very simple procurement or see as example in general, it takes.
[00:18:57] Around 60 steps [00:19:00] from you have an idea until you get that product delivered to you. And there are going to be a analyst, back and forth conversations, different time zones, different cultures and languages in there, that AI can help a lot for both party, buyer and supplier. So we believe that the B2B will go to a two A.
[00:19:24] Agent to agent, that will be the future.
[00:19:27] Mike Kaput: So I love that example. There's 60 full steps in that process.
[00:19:32] Mike Kaput: Can you, I'd love to hear maybe another example or two from you, just concrete ways. How are entrepreneurs and small businesses using something like Accio work today? I think that really hammers home for people just how much friction there is in this process that agents can help with.
[00:19:49] Kuo Zhang: Okay, sure. So, we actually shoot a video for that. If you have interest, send
[00:19:56] Mike Kaput: Yes.
[00:19:57] Kuo Zhang: Send to you. So just describe a [00:20:00] kind of, if you are a guy who's leveraging global supply chain assets, SMB. So in general, what's your daily life look like? So you have some ideas. For example, if you want to source a mountain back.
[00:20:18] And resell in UK or in us. And then you need to describe what do you need like in email or in rq, and then you copy paste This require to about a hundred or 20 suppliers. With your requirement back and forth. And don't mess up with name. Okay, and then you just wait for that reply, some fast, some slow, and then you start manage Excel to compare all the columns from MOQ to [00:21:00] price, to certificate, to quality to weight, and you can name it and that kind of conversation back and forth probably for a week until you gather a spreadsheet with dozens of replies of suppliers.
[00:21:15] And then you select probably two or three or five of them, you start to further negotiate for further terms. And then you need to talk with them what, how to kind of, manage the inco terms, the deliveries, how to pay for them. And then you gather product you have back and forth on the examples. And these are the kind of, daily work life of SMB who is sourcing, let's say overseas.
[00:21:45] And in general, that takes for weeks or sometimes for months, you just starting from an idea and then find a right guy who can make it for you and then get the results. So with the help [00:22:00] of ai, actually what do you need to do is you tell the agent what your needs look like and what you are kind of, looking for.
[00:22:10] What's your budget look like?
[00:22:12] Kuo Zhang: And then the agent start. Looking for the suppliers who specifically match your needs and help you to communicate back and forth with the suppliers and they can generate the projects for you with all the columns that you need, probably just within a day if you are lucky in hours and starting from there that you can start to push forward other communications and the agent can help you along the way To finish.
[00:22:42] It's a kind of contract in the end. So with the red MOQ red price and with the red shipment method, Inco terms and payment terms. And then we can leverage Trade Assurance as actual B2B payment help you to protect the entire [00:23:00] order. So that kind of streamline the B2B business a lot of easier than before.
[00:23:06] So I just using sourcing as example, there are many other examples as well. So you can imagine that for SMB in general who don't have the global source experience,
[00:23:19] Kuo Zhang: very limited time and resources and don't understand about the terms. And many of cases, they don't know how to write a design pack for your product, right?
[00:23:31] The agent can help you to handle all that parts and you only focusing on the most important question, like what problem you are going to tackle. Who are your customers, unless agent, to help you to manage the rest?
[00:23:46] Mike Kaput: I love those examples. That just shows there are so much, so much opportunity here to dramatically streamline and speed up the trade process, especially in B2B with all these tasks that frankly, [00:24:00] like businesses, entrepreneurs, even SMBs or people who are very expert at sourcing, nobody wants to be spending.
[00:24:07] Hours, days, weeks, on these tasks, they want to get back to the business of selling. So it sounds like agents can really streamline that process.
[00:24:15] Mike Kaput: I'm curious at a high level, with whatever you can share, what kind of technology is powering Alibaba.com's AI transformation, including Accio work? We've talked about agents kind of broadly.
[00:24:27] Can you maybe go into a little more detail there?
[00:24:31] Kuo Zhang: Okay, we have a formula actually in general to describe what the agent or Accio is, is composed of. So agent = 2 x model x harness x context.
[00:24:47] Kuo Zhang: So why is the multiply? Because each element is zero. The entire equation is zero. The model actually is,
[00:24:56] How to say the model includes a lot is, [00:25:00] intelligence, power, and we are using, we're leveraging all the frontier models in the world, including the queen. And also we do some post train on the queen, small size, model to make sure that it can, has the same power of the big size frontier model because of the post training that we use.
[00:25:20] And also we do a lot of serving stuff as well. To make sure that we help. Better cash rate and we have a better performance of that model. And the harness actually is the key as well. So harness contain all this, logic, the loops, the gut rails, the tools that you need to contact, connect, like we manage to, help the users to manage many platforms.
[00:25:50] Then that you need to, to have tools to connect to all these platforms, to aggregate data, help you to upload listings to the market, research, so on and so [00:26:00] forth. And the last but not least is the context, the con context, including all the real business data, the conversations you have, and also the data from platforms.
[00:26:14] The broader the data is. It's a kind of, more, broader view of full understanding of the, background and the trend and the category. Actually you can have a better decisions and result out of that. So this is the kind of the formula that we are composed and we kind of keep improving on to get the better, agents and to make sure that.
[00:26:43] We are continually improving this system. We are invented a bench we call commerce agent bench.
[00:26:53] Kuo Zhang: So, you know, there are all kinds of bench, including mathematics, coding, so on and so forth. But [00:27:00] when we talk about the business environment, a little bit messier than all the bench before. And we, come, we are kind of, distill from millions of conversations.
[00:27:15] I mean the real conversation between the buyer and the sellers and different execution workflows as well. And we abstract 107 different tasks
[00:27:25] Kuo Zhang: To make sure we cover all the categories where people are doing business on platform and we using this bench. Kind of to test all the result, using different models and leverage our harness and using the real context of the real business to make sure that we can evaluate the result actually is measurable is, reliable and the most important is affordable to all the s.
[00:27:56] Mike Kaput: I really like that framing. That's an elegant way to put it. So it sounds [00:28:00] like, you know, you have your model, which is, you know, doing various different types of post training on some version of Alibaba's Qwen family of models, which are extremely formidable. That's kind of your raw intelligence multiplied by a harness, which is kind of the, you know, proprietary secret sauce, bit of code, logic, guardrails, essentially all the.
[00:28:22] Machinery or architecture around the model to help it be AGI agentic in the way you want, and then multiplied by the context, all the business data, the domain expertise, Alibaba has, et cetera. That's a really powerful mix, and it is why you cannot just duplicate something like this in a different business or a different context.
[00:28:42] There's these three elements, it sounds like, working together to make all of this possible.
[00:28:48] Kuo Zhang: Exactly.
[00:28:50] Mike Kaput: So what we've talked about is amazing. I mean, there's really cool results by saying, okay, applying agents to the e-commerce process, especially a B2B [00:29:00] process through alibaba.com is producing awesome results and things for customers, for buyers, for sellers.
[00:29:06] But this also, you kind of alluded to this when you said a to a like agent to agent commerce, like stepping back a bit here, how do you think AI agents are actually going to be transforming. Commerce at large. It seems like there's some big implications here. how is this changing Alibaba dot com's own business?
[00:29:25] Kuo Zhang: So the first of all, we, we think of, in the future what is going to look like. So we think a to a is the answer. Since the both party, I mean, the suppliers and the buyers both want to save their time and get efficient, result. as much as possible. So there are needs and also there are network effect.
[00:29:54] So then you get a network of suppliers and you have a net network of, [00:30:00] fulfillment, supply chains. And you have all demands in the world. You know that every small and medium business all over the world, when you ask them, do you need global supply chain? Do you need kind of a agent team to help you to save time?
[00:30:18] I think it's a yes for all the companies that we we meet and why they do not using the global supply chain or do global business is because of their size, power, limited of resources. If we, we can kind of, give them this power to access, I think every small SMBs can, unleash their full power. Since they understand the local business, they understand the local needs, and that means that we can have a more broader audience of our customers, the users.
[00:30:55] So always comes first from the demand side. So when the demand side [00:31:00] is getting larger and larger, we can kind of aggregate the supply side all over the world as well. So it's all the suppliers, they want to get more business. And the more that they review actually encrypt by agent or provide their real business data, provide their real kind of manufacturer capabilities that is more easier for them to connect to this network.
[00:31:25] And when the agent to agent kind of, working together is more like a multiply, it is not like a search engine. I recall some of the results, for example, I have. Hundreds of thousands of suppliers, I only give you 1000 and using this kind of ranking mechanism and you can only talk to dozens of them since you don't have enough time.
[00:31:49] But when the A two A network work together, it's easily help you to kind of map with a much huge network and it's much easier for you [00:32:00] to kind of get the best match with your requirement, with your budget, with your timing. And it's much easier for us to kind of, get a better, result out of the demand and supply, requirements in the world.
[00:32:14] So I think in, if we think about that way, that is much bigger than what we are doing today using the search engine like business.
[00:32:26] Mike Kaput: So. The technology is really one part of a transformation like this, both in the industry and with what you're enabling for customers. I'm curious about looking at alibaba.com, like what had to change inside the company?
[00:32:43] As you move from experimenting with AI to building and operating products and workflows entirely around agents, that probably has some implications for how your company operates, how talent works. I'm curious if you could unpack that for us.
[00:32:59] Kuo Zhang: sure. So [00:33:00] I just tried to, tell you as much as possible, but the first thing that Accio is operates as a startup.
[00:33:10] Team or company within Alibaba Group.
[00:33:13] Mike Kaput: Okay.
[00:33:14] Kuo Zhang: It's just, you can, you can imagine that alibaba.com has its own pace, has its own business model, and also since we are a public company that we, it has this own plans as well. We'll ask you or ask you, work is a completely different creature that, for example, oliver.com has a feature update biweekly.
[00:33:39] Or by month since that system already exists for 26 years, as you mentioned, it started from 1999. So every feature that you add on, you need to be compatible with the existing features and does not impact the customers who is not familiar with you. Do need to have a lot of, AB [00:34:00] test and have a lot of kind zone off for any future goal line.
[00:34:04] Will Accio and Accio Work. Is it has a version by a daily basis, so every day it has a new version.
[00:34:14] Mike Kaput: Oh, wow.
[00:34:15] Kuo Zhang: Yeah. Every morning that I get up, the first thing is that I upgrade this software or this agent to see what's new there. Since the agent, actually many of the features they read themself.
[00:34:28] Kuo Zhang: It gathers all the information, the requirements from all this kind of, daily execution environments.
[00:34:35] And it's ranking actually. People help to add ranking and fix box and kind of a lot of coding is, written by themself. So it's just evolve in a different pace. And also what we need to integrate two together is think about a audible.com. It's part of harness of this agent, right? We need to get this [00:35:00] website first.
[00:35:02] To get it all this, services, cly, it can be, invoke, this can be, used, leveraged by this agent easily. Okay? So it is not only designed for human, but it is also designed for agent as well. And this two, two toolkit need to work simultaneously. And now we are kind of integrating. Back to aaba.com.
[00:35:30] We call it the AI mode. Okay. Meaning that all the users can use aaba.com as an agent for them. So it, so we started from a startup team to, to kind of create this business. And also we, we are kind of, using the, existing engineering team to kind of, give the whole, platform a upgrade. That so that it easy to be used by the agent.
[00:35:59] And now we are [00:36:00] integrating the two product together. So all the users on a audible.com easy to assess to the to Accio as well. So it's a step by step approach within alibaba.com.
[00:36:13] Mike Kaput: That's really fascinating. So it sounds like Accio work is almost like its own innovation hub or lab or business within alibaba.com, where it sounds like alibaba.com, because rightly so, of the legacy and the history with customers and the need for continuity.
[00:36:31] Shipping updates at a biweekly or even monthly rate, whereas Accio work is literally updating almost every day or a version every day. It sounds like it is also updating itself. And then you had mentioned CLI or Command line interface to then it's actually starting to learn, or you're implementing having it work.
[00:36:53] To talk back to alibaba.com Right. To start integrating there so you can incorporate AGI [00:37:00] agentic features more into the main business. Do I have that right? That that's incredible.
[00:37:05] Kuo Zhang: Yes. And today since, Accio is integrated in alibaba.com as AI mode now, alibaba.com is upgrading daily basis as well.
[00:37:16] Mike Kaput: Hmm.
[00:37:16] Okay. So it's made it possible to then ship faster based on what AI and agents are now enabling.
[00:37:24] Kuo Zhang: Yeah, exactly.
[00:37:26] Mike Kaput: So a couple more questions here for you Kuo. As we wrap up here. I'm curious what, as you've gone through this incredible, journey with Accio work and alibaba.com generally, what has been hardest or messiest about going from all these AI experiments to actually building agent-based products and operations?
[00:37:49] Anything jump out to you as a particularly big obstacle? Yeah,
[00:37:54] Kuo Zhang: so I think the biggest challenge is everything is just [00:38:00] moving so fast, right? Previously, it's easy for us to predict what this look like for three years. Like you have a three years business plan and you execute the quarterback quarter.
[00:38:14] But now in the AI era, three years means forever. So after three years, actually nobody can, can predict what will happen. So think about three years ago, nobody can understand or can predict. The AI actually can penetrate coding. Let's say for 95% probably. I did not see any coder. Now don't use any kind of ai.
[00:38:42] Mike Kaput: Yeah,
[00:38:42] Kuo Zhang: to do that coding in daily basis. You can imagine that in all different business categories, the penetration will be same or even higher. If we find out a way that using agent [00:39:00] team is far more efficient than the way that you do business before, I think the adoption rate or speed going to be faster than ever.
[00:39:11] And also the technology is evolving. Faster than we expected as well from the hardwares to the models, to all this kind of agent design systems. If you, if you ask me the formula of a agent, let's say six months ago, probably, I cannot answer that question. So now more, more and more we understand what's the kind of underneath technology were the key factors and how we measure the success of the agent.
[00:39:43] Actually, we now can answer your questions and if I have this knowledge, let's say travel back to six months ago, a lot of things that we can do differently. But yeah, we just, we don't know what we don't know. And now I think there's a lot of things [00:40:00] that we don't know yet, and we just need to keep, keep learning and keep on the pace of this kind of AI era to make sure that the thing that, I think what we do right is we found the right problem to solve, right?
[00:40:18] If the problem is not a real problem. It's a fake problem, then no matter what do you, what do you do? Actually in the end, it is not going to get any business impact. If it is a real business problem and they can solve a lot of, issues for those SMBs, or we think that there are going to be a huge potential of this AI that is not unleashed yet.
[00:40:45] Mike Kaput: That's incredible. I love how you've broken that down.
[00:40:47] Mike Kaput: So you know, final question for you, Kuo. I'm curious, I want to circle back to something you mentioned. You talked a bit about Commerce Agent Bench, and I'm curious, can you talk a little bit more about this? Because the kind of issue here is as agents actually start executing real business tasks across knowledge work, not just commerce.
[00:41:08] I'm curious, alibaba.com has this kind of interesting, unique approach for measuring how or whether or not agents actually get the job done. So Commerce Agent Bench's this open sourcing benchmark that you had mentioned for evaluating AI agents. Can you just tell us a little more about that and we'll kind of wrap up here just talking about how, Commerce Agent Bench makes it easier to understand, are agents doing the right work and are they performing well?
[00:41:39] Kuo Zhang: Yeah, sure. It's, so the, if you are familiar with this idea of Pareto Frontier
[00:41:46] Mike Kaput: Yep.
[00:41:47] Kuo Zhang: Now Pareto Frontier. Yep.
[00:41:48] Kuo Zhang: Yeah. Everybody's talking about that. So think about this is a new system. It's a kind of new framework for the commerce environment of Pareto [00:42:00] Frontier. So you use this model to solve mathematics is perfect.
[00:42:06] When you use this model to solve, let's say, the real problems in the business world, many of the cases is failed. So this kind of a commerce workbench is just for you to narrow the real business scenarios with real business case and real business context. And this the real harness, which is Accio. And using that kind of, test.
[00:42:33] We have 107 task yet, and we, that numbers keep evolving. If, next year you ask me, probably it's going to be 300 and more and we just. Keep involving this, this work bench to make sure that we can capture more and more complex, complicated, problem in the world. And now the, we, we actually see it in different [00:43:00] dimension.
[00:43:00] The first is a pass rate to make sure that this agent can perform in each of the scenarios. If it cannot deliver, all the result is wrong. The pass rate is zero, it's only one or zero. And then we also consider about the time, how much time and steps that the agent can give the result with a certain model.
[00:43:24] And also we consider about the price. So at what price that you solve these problems. So in the end, only the affordable AI is a useful AI for SMBs. So you will have the best model, but it's not affordable to do daily basis work. It's not useful as well. So we are, we are kind of, redesigning the framework, thinking from the business perspective for SMBs to make sure that we are routing to the best model fit for the right problem.
[00:43:58] So that's, [00:44:00] that's why we are kind of, inventing this and we are not only can kind of competing, or comparing model with model in the same harness. Like we put, we put all the frontier models including the Qwen model in a harness of the Accio and you can compare which model solve which problem more efficiently.
[00:44:24] And we can also use this to, We have a benchmark for the agent as well. For example, with the same 107 tasks, we are using Accio. We are using Codex and we're using Claude Code to solve that 107 same task. Accio actually can have a higher pass rate, this less than half of the price. The Meanings SMBs with the same budget.
[00:44:52] You can actually do twice of the job. So that's an another way to see how we can leverage this [00:45:00] work, work bench as well. So we're just, understand which model fit for which problems, and we can help the SMBs routing through different models to solve different problems and get the best result.
[00:45:16] Mike Kaput: I love that.
[00:45:17] And just to kind of reiterate here as we wrap up, so when we talk about the Pareto Frontier, this is basically we're just to kind of clarify for anyone who hasn't heard that term, or if you wanna connect the dots here, that's when we're really just trying to determine what are the best trade-offs among different qualities of an AI model, right?
[00:45:35] Like we all have encountered this in one way or another. You wanna go use, say, GPT-6, Astra, or Fable 5 or something. It is really good at certain things. It's bad at others, and it costs a lot of money. There's different trade-offs, right? It could be slower in some ways or faster in others. That's exactly what we're talking about here.
[00:45:54] So Accio work, the reason this benchmark matters is because you actually sound like you have a way [00:46:00] to actually look at over a hundred different tasks that are actually. Important to businesses and you're able to benchmark the agent and on those exact tasks from a Pareto Frontier perspective. So it is the best trade off between speed, reliability, price, performance for those tasks.
[00:46:22] Do I have that correct?
[00:46:23] Kuo Zhang: Exactly.
[00:46:25] Mike Kaput: Okay, great. So that, and that's just for everyone listening, that is kind of exactly what we're getting at is what's so important with these types of developments as we move into this agentic era is the ability to actually say conclusively, can this agent actually do the work?
[00:46:41] So Kuo, I thank you so much for sharing Alibaba.com's journey. For sharing all about Accio work. This is, honestly, feels like looking into the future. I can't wait to see how agent to agent commerce evolves. Thank you again for being here and for sharing Alibaba dot [00:47:00] com's AI transformation story.
[00:47:02] Kuo Zhang: Thank you for having me.
[00:47:03] Mike Kaput: Thanks for listening to our AI transformation series from the Artificial Intelligence show to keep learning, visit SmarterX.ai where you'll find on-demand courses, upcoming classes, and practical resources to guide your AI journey.