Skip to content
Alibaba.com President on Accio AI, Global Ecommerce & "A2A" Business artwork

Alibaba.com President on Accio AI, Global Ecommerce & "A2A" Business

Schwab Network

September 10, 2026

Kuo Zhang, president of Alibaba.com, joins Market on Close to explain how the AI buildout is impacting the global commerce space, especially for smaller businesses. He also offers insight into Alibaba's Accio AI agent and ways the company balances tech costs with efficiency.

Speakers Kuo Zhang

TopicsInvestingBusiness

SPEAKER_1 (0:00)

Joining me now to discuss the role of AI in the future of e-commerce is Kuo Zhang, the president of alibaba.com. Kuo, thank you so much for being with us today. Let's start broadly here. I mean, AI is rapidly changing the way that businesses operate. As we look more specifically at small businesses, how is it affecting how they're actually getting launched and how they're operating, particularly for entrepreneurs who don't have large teams or large amounts of resources when they start?

Kuo Zhang (0:31)

You said a very good point, that what we think AI today is a kind of agenda team that you can delegate your work to, and what it most can help for the SMBs, who would like to access the capability, the supply chain that normally only the big companies can have. Accio actually is a tool to help SMBs from sourcing from globally and also manage their work front, their store front, manage the business.

Accio actually is the first AI workspace for e-commerce. That is how we designed it.

SPEAKER_1 (1:10)

So Accio was designed to help with pretty much everything. From market research to product development, I watched your videos about it and it looks incredibly capable. Sourcing, logistics, how much of the traditional work of starting a business, can AI now realistically take off of an entrepreneur's plate?

Kuo Zhang (1:31)

It's a good question. So what we are focusing on more are the repetitive works. The repetitive works, for example, you talk about the sourcing, so they need to kind of starting from idea and generate the design packs, talk to dozens of suppliers, and then select one.

So in general, it takes for months. Now with the help of AI, it can happen in hours or days.

And the seller needs to kind of sell the product in multiple platforms, and they need to manage the listings in different platforms and manage the customer support, and also manage the data. So did I make money or not? And these repetitive works can be handled by AI as well. So those are the things that we think AI can help the most.

SPEAKER_1 (2:19)

And you just released a benchmark at CoCreate, showing that Accio actually completed 107 commerce tasks at more than 50 percent lower the estimated cost of OpenAI's Codex and of Anthropics' Claude code. What is allowing you to deliver comparable performance at a fraction of the cost?

Kuo Zhang (2:40)

Okay, that's a great question.

So we have a formula. The agent equals to model, multiply harness and multiply context. And for models, actually, we are defining or we are routing different tasks, different jobs to different models to make sure that we use the best model and the most cost-efficient model to solve the right questions. And also, we did a lot of tools, a lot of context management to make sure that we can deliver the reliable result with the most meaningful cost. We think that affordable AI is a useful AI, that SMBs don't have too much budget on that, and they rely on the agent to do the daily work. And that's why we spend a huge effort to do that.

SPEAKER_1 (3:31)

And one of the things I noticed in your CoCreate presentation about using Accio and some of the capabilities here, is that it can actually take work off your plate, like negotiating with the suppliers. At the end, the human steps in and makes sure to make the agreement. But as we move more into AI doing processes like that, giving, from giving recommendations to actually doing the negotiation with suppliers, calculating landed costs, booking shipments, if there's a mistake, that would become much costlier. So how do you make sure that businesses can trust the AI agent to take these actions?

Kuo Zhang (4:11)

So most important part, we think, is set the boundaries.

So the example you gave is that in the final step, you select which partners you work with. So that part that we cannot let AI to decide for you, you need to make final decisions. When you make the payment, actually, it's the human that need to make sure these are the terms, the contract and the price you agreed on, and then you can make decisions whether you take it or not. And these critical decisions that need to be happen to human, it's not that the AI can take everything from your plate. That is how we deal with it.

SPEAKER_1 (4:48)

And how does a small business go about measuring the ROI on an AI agent when they're looking at the overall efficiency?

Is it time saved? Is it at lower sourcing costs, higher margins, increased sales, or is it just ultimately that your business can operate with fewer employees?

5 more minutes of transcript below

Thousands of transcripts fetched by people building searchable podcast archives

Fetch the whole transcript

The demo key returns a sample episode in full, no card needed:

request
curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Markdown with the speakers named, for your notes, your knowledge base, or anything that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

request
curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000788912190