Does Your Cross-Border Store Actually Know Your Customers?

Let me ask you a question first.

Say you run a cross-border e-commerce store. Open your admin dashboard right now — can you answer this: yesterday, someone added your product to their cart, then closed the page right before paying. Who was that person?

What's their name? Where did they come from? What else did they look at in your store? And why, in the end, didn't they buy?

My guess: most sellers can't answer.

It's not that there's no data. There's plenty of it: one copy in the ads platform, one in the store backend, one in the customer service chat logs, another in the email system — and finance may still have an Excel file lying around.

But none of them talk to each other. You're holding a pile of numbers, and you still can't piece together one complete human being.

This reminds me of a talk from September 2024. The speaker was Dinh Le Dat, co-founder of the marketing technology company Antsomi, whose product CDP 365 does exactly this: pull customer data into one ledger, then run marketing with AI. He was speaking at a cross-border e-commerce forum in Vietnam put on by UOB FinLab, on how AI marketing applies to cross-border business.

What he said happens to answer the question I opened with. Let me walk you through the core of it.

1. AI Isn't the Future — It's a Tool Already on the Job

For many people, AI marketing still sounds like hype.

Dat's take is refreshingly grounded: AI arrived a while back and is already doing real work — two things, in fact. One is automating personalized experiences. The other is helping people make data-driven decisions.

What does automating personalized experiences mean?

Think about it. In a cross-border store, what shows on the homepage for whom, how search results are ranked, what recommendations sit on the product detail page — all of this used to be configured by hand. One version per market, ten versions for ten markets, until you burn out.

What AI does is let the system tailor the experience to each person, on its own. Everyone who opens your store sees a version tuned just for them.

Cross-border business needs this far more desperately than local business. You're facing several markets at once — different languages, different habits, different taboos — and you can't hire someone who understands both the local market and store operations in every single country.

AI is the one "employee" that can understand every market at once.

But first, it has to get to know your customers. And getting to know them runs on data.

2. One Master Ledger, and the Three Things It Strings Together

That day, Dat made a call: from 2024 to 2026, competition in global retail would revolve around three things — customer experience, customer data, and personalization.

Looking back now, that prediction doesn't look far-fetched.

Three things, but really one chain: experiences need personalization, personalization runs on AI, and AI feeds on data. At the very bottom of the chain sits data.

So how do you build up that data? This is where a term comes in: CDP, customer data platform.

Simply put, it keeps one master ledger for every customer. What they clicked in your ads, what they browsed in your store, what they bought, what they returned, what they complained about to support — all of it recorded under the same individual, adding up to one complete portrait.

Without this ledger, your customer is a pile of numbers; with it, a living, breathing person.

Only once AI knows this living person can it even begin to serve them.

3. Three Places That Directly Make Money

Enough theory — let's get practical. He pinned AI marketing's use cases to three places that can directly move revenue. I'll go through them one by one, and run the numbers with you as we go.

First, conversion-rate optimization.

From walking in the door to paying, a shopper passes four spots: the homepage, the product listings, the product detail page, and the search bar. AI recommendations can stand guard at every one.

The search bar is worth the most. Someone willing to type a search comes with intent already in hand. When AI tunes the results, what comes back is what they're most likely to buy. And just like that, conversion climbs.

Next, pulling back the people who are about to leave.

The research firm Baymard Institute has long tracked one number: on average, roughly 70 percent of shopping carts are abandoned before checkout.

Seventy percent. More than half of your ad spend actually goes toward helping customers "get to the payment page."

AI has two plays here. One: at the moment a user is about to close the page, push a coupon or a recommendation they'd actually want, and intercept them. Two: abandoned-cart recovery — automatically send reminders and offers to people who've already left, and win those lost carts back. The machines run the whole show; no need to hire anyone to watch it.

Finally, lifecycle marketing. This is the one most worth doing the math on.

A lot of owners fixate on a single number: how much we sold this month. But what really decides life or death is customer lifetime value, or CLV.

Say you spend $100 on ads to bring in a customer. They buy once, gross margin $80 — this deal loses money, and the more you pour in, the more you lose. But what if they come back five times a year, and each time the AI, knowing their history, recommends a little more precisely than the last?

Same customer: in the first case you're out $20; in the second, you make several hundred.

The difference between the two is that ledger. Acquisition cost, cost per order, engagement rate, lifetime value — only when all four numbers sit in one system where you can see them whole can you actually act on them.

On its website, Antsomi features the case of the coffee chain ZUS Coffee: after adopting CDP 365, the number of customers converted grew 3x, and revenue grew 6x.

Those are the numbers they put up themselves — take them with as many grains of salt as you like. But the direction, I believe, is right.

Finally, Back to That Person From the Start

The person who added to cart and walked away will still go buy from some other store.

Your store didn't keep them — not because your products weren't good, but because your store didn't know them. The store next door that has its data and AI working? It knew what that shopper wanted the moment before they closed the page.

So don't rush to ask what AI can do for you. First pull your customer data into one ledger, and let the machine get to know every person who has ever come through your door.

AI is the amplifier. Data is the electricity. Without power, the amplifier is just a nice-looking lump of iron.

Here's to smooth sailing as you go global — and may every cent of ad spend land on a person you can actually see.