When Generative AI Crashes Into Cross-Border E-Commerce, Should Small and Mid-Sized Sellers Follow?

A while back, a few old friends of mine who run cross-border e-commerce businesses invited me to dinner.

A few rounds in, one of them suddenly put down his glass and asked me: "Run, be honest with me. I've got a dozen or so people at my company, and every day it's the same grind — uploading listings, writing product detail pages, shooting short videos. These past two years, AI tools have been popping up like crazy: ChatGPT, Midjourney, and the platforms are building their own assistants too. Should I use them or not? If I do, I'm afraid of stepping on landmines. If I don't, I'm afraid of falling behind."

I didn't answer right away.

Because I knew this wasn't just his dilemma. It's the same wall that more than a hundred thousand cross-border e-commerce companies and over a million people working in this industry across China have collectively run into over the past two years.

First, a fact: this industry has quietly grown tenfold

What does "running into a wall" mean?

It means that while you're still debating "should I get into this business," the industry's scale has long since surpassed your imagination.

Let me give you a few numbers.

Over the past five years, China's cross-border e-commerce import and export volume has grown tenfold. In 2023 alone, it reached 2.38 trillion yuan, up 15.6% year over year. Of that, exports accounted for 1.83 trillion yuan — growth of nearly 20%.

What does that mean in practice?

China now has more than 120,000 cross-border e-commerce companies, over 200,000 independent sites (i.e., seller-owned D2C stores outside the big marketplaces), more than 1,000 cross-border e-commerce industrial parks, and more than 2,500 overseas warehouses adding up to over 30 million square meters. The government has approved 165 comprehensive pilot zones in one go.

Can you still call this a "small business" with a straight face?

The more uncomfortable part comes next. This huge crowd of sellers is no longer staring only at big platforms like Amazon and Alibaba.com. In 2024, the independent site market reached 3.4 trillion yuan, accounting for 35% of the entire B2C market. One brand that started out selling power banks posted 71.75% revenue growth on its independent site in 2022.

While everyone else is still fighting for traffic inside the platforms, it has turned itself into the traffic.

Livestream selling has barged in too. Short-video apps have been the world's most-downloaded app category since 2020, and cross-border livestreaming has become the new hot spot. Sellers used to work trade shows, grinding out one customer at a time; now a single livestream can close a whole wave of deals at once.

Second fact: generative AI arrived faster than you think

Cross-border e-commerce is surging, and generative AI is moving even faster.

What is generative AI?

In plain language: you give it one sentence, and it hands you back a piece of copy, an image, a video, even a line of code. ChatGPT counts as one. So do the image-generation tools. And Amazon's conversational shopping assistant Rufus counts as one too.

It's different from the earlier generation of AI that "could only classify and predict." This generation can "create."

The pace of adoption in China is a little scary.

AIGC (AI-generated content) applications have surpassed 73.8 million users — an eightfold increase in one year. Some experts predict that by 2035, generative AI will create nearly 90 trillion yuan in economic value worldwide, with China capturing more than 30 trillion yuan of that — over 40%.

Regulators are laying track too. In August 2023, the Interim Measures for the Administration of Generative AI Services took effect — encouraging innovation with one hand while setting rules with the other. By April 2024, 117 large models in China had completed service registration.

This isn't the future. This has already happened.

What is AI actually doing in cross-border e-commerce?

Don't rush. Let's take it apart.

Right now, AI's work in cross-border e-commerce comes down to three things: personalized recommendations, insight mining, and content production.

Sound abstract? A few scenarios and you'll get it.

There's a foreign-trade sales rep who used to spend an entire afternoon squeezing out a single prospecting email for a new client. Now he drops the company name and product name into a tool, tweaks a few words, and a respectable prospecting email comes out.

Amazon launched an AI shopping assistant called Rufus. When overseas buyers ask, "Is this good for camping?" or "What's the difference between these two?" it recommends and compares right there in the chat box.

Then there are platforms like Alibaba.com, which rolled out an "AI Business Assistant." The moment a buyer sends an inquiry, it instantly grasps the intent, auto-generates a reply draft, and even fills in the quotation for you.

Product detail pages, product images, livestream scripts, inventory forecasts… AI has seeped into every crack of cross-border e-commerce.

On Alibaba.com alone, AI-generated product listings have surpassed 100 million, and merchants can use AI in more than 40 scenarios.

Think about it: if a tool can take over half the work of copywriting, design, and customer service, do you dare not use it?

But hold on — don't charge in just yet

After that dinner ended, I had several deeper conversations with these friends, and I noticed something interesting.

Lots of people say AI is great. Very few have actually made it work smoothly.

I counted the pitfalls that almost every small and mid-sized seller has stepped in. There are exactly four.

Pitfall one: everything AI writes tastes the same.

The reason isn't complicated. One company lists the same product across several platforms, which should call for several versions of the copy. But everyone uses roughly the same models, feeds them roughly the same data, and some people just stick with a single tool.

The result? The same product's description on Amazon, on the independent site, and in the livestream room looks like it was carved from the same mold.

Consumers feel fresh the first time they see it; by the third time, they're annoyed. Brand distinctiveness is gone, and loyalty goes with it.

Homogenization is AI's identical gift to every lazy person.

Pitfall two: not enough people, and they don't know how to use it.

New AI tools arrive by the batch every month, with feature updates moving as fast as phone OS upgrades. Small and mid-sized enterprises were already short on cross-border e-commerce talent, and now they have to chase another AI course on top. Worse, many bosses themselves aren't sure which tool to pick, let alone how to connect it to the business. Employees have no one to lead them, tools have no one to teach them — the whole company is feeling around in the dark.

Pitfall three: AI sounds perfectly convincing, and you have no way to verify whether it's right.

AI has a flaw: it doesn't guarantee that what it says is correct. It can talk nonsense with a perfectly straight face.

Yet small and mid-sized enterprises are precisely the ones short on data and case studies to verify whether AI's marketing advice actually works. Verification costs money, time, and people — and cross-border e-commerce sales costs are already high. Nobody wants to pay an extra "trial-and-error fee."

The most common scene: mid-level and frontline employees are the most enthusiastic users, but they're also the least equipped to judge the quality of the output. The people using it and the people who can judge it are not the same group of people.

Pitfall four: feeding your data to AI is like handing your lifeline out the door.

For convenience, employees dump piles of customer purchase records, new product specs, and marketing strategies into AI tools for analysis. There are 100-plus large models on the market; today you use this one, tomorrow you switch to that one.

Once sensitive information leaks into a competitor's hands, your product plans, client lists, and pricing strategy are laid completely bare.

You treat AI as a helper; it can turn into a moving company at any moment and carry the most valuable things in your house out the door.

So what should small and mid-sized sellers actually do?

I told those friends: don't panic. The pitfalls are real, and so is the path.

I took what I saw in them, added what I've thought through myself, and boiled it down to five points.

First, squeeze the free stuff dry before talking about anything advanced.

If resources are tight, don't start by throwing money at paid tools. The AI tools and companion courses built into platforms like Amazon and Alibaba.com are free in many cases — if you've never touched AI, start there. If employees pay for tools out of their own pockets, reimburse them within a reasonable range to encourage learning. Attend platform skill competitions and industry meetups whenever you can. You can also partner with universities and use their training software to run marketing simulations — practice first without burning real budget.

If you can learn it for free, don't walk into the paid casino.

Second, don't fully trust AI — build a verification mechanism.

Someone must stand guard over AI-generated content. How? I'll give you three handles.

One, use big data. Cross-border e-commerce platforms are sitting on oceans of consumer data. First dig out real consumption habits, then have AI generate content based on those insights, and finally loop the data back in to verify the results.

Two, ask the buyers. Wherever your goods are sold, find partners in that country and ask how the local version of your marketing content reads to people there. Frontline feedback beats guessing.

Three, run war games. University training bases and livestream bases can both be used for practice. Test in a safe environment first, then take it to the real, live market.

Third, find partners and travel together — don't go it alone.

Cross-border e-commerce isn't something one person can carry. Platforms, technology vendors, and industry peers can all become your partners. Form strategic partnerships to share resources, technology, and experience. Industry communities are full of people studying "how to actually apply AI to cross-border marketing," and the lessons they've earned by stepping in pits are worth mining. The government is also promoting public training-data platforms, and companies should actively participate and be willing to put their own data on the table. An industry moves forward never by one or two heroes, but by a crowd of people willing to blaze the trail together.

Fourth, give AI a job description — don't let it cross the line.

AI can do a lot, but it can't do everything. The company must draw clear lines: which work goes to AI, and which work must be done by humans. AI can draft livestream scripts and detail-page copy, but the final polish and tone must be supervised by a person, or everything comes out looking the same. As AI's capabilities change, your evaluation standards must change with them. I'd even suggest writing AI tools into your workforce plan as a "virtual employee": with a position, with clear responsibility boundaries, and with regular performance reviews.

Fifth, when it comes to law and data security, not a single line can be crossed.

Before adopting an AI tool, confirm it has completed service registration. While using it, follow the rules one by one: consumer privacy protection, intellectual property, fair competition. Ideally, appoint dedicated legal counsel or hire a professional lawyer to run regular compliance reviews of the company's marketing activities and AI usage. At the same time, manage the inside: know what your employees are uploading online.

Compliance isn't a shackle — it's the safety rope that lets you travel farther.

Final words

At the end of that dinner, my friend was silent for a moment, then said:

"The way you put it, AI isn't a question of whether to use it anymore. It's a question of how."

Exactly.

What generative AI brings to cross-border e-commerce is a genuine opportunity. It's especially big for companies that came over from foreign-trade factories, that run several platforms at once, and that also have to do livestreaming on the side.

But opportunity has never equaled results. Pour AI into your business like a cure-all, and it becomes a landmine buried under your business; treat it as a knife that needs a human hand on the handle, and it can help you cleave open new markets.

Tools will change. Industries will change. Only one thing stays the same: in the end, the winners are always the people who both dare to use new tools and remain clear-eyed about where the tools' limits lie.

May you and your company claim your own spot in this wave.