Ask an AI how to market a bakery and you’ll get an excellent answer. It will mention Instagram, loyalty cards, seasonal products and local SEO.
It’s the kind of answer a smart marketing student gives. Correct in general. Useless in particular.
Where AI’s knowledge comes from
Large language models learn from enormous amounts of public text: websites, books, articles, forums. That is why they know so much about marketing in general.
But think about what is not in that public text:
- The Zalo message where your customer said she’d stopped coming because parking got difficult.
- The fact that your Tết orders always arrive three weeks early, from the same corporate buyers.
- The rumour in your street that a chain store is opening next month.
- The real reason your best salesperson closes twice as many deals as the others.
None of this is published anywhere. So no AI has learned it. And this is exactly the information that decides whether your marketing works.
The general-knowledge trap
Much of the marketing advice on the internet, and therefore much of what AI has learned, describes markets very different from yours. Different payment habits. Different trust signals. Different platforms.
In Vietnam, a business might win through Zalo, Google Maps reviews, word of mouth in a residential compound, or a respected person in the community. The global playbook often puts these near the bottom, if it mentions them at all.
When you follow general advice, you compete on the same ground as everyone else following the same advice. The ground where you can actually win is local, specific and unpublished.
Your data is your moat
Here is the twist. The more AI spreads, the more valuable your private, first-hand knowledge becomes.
Everyone has access to the same models. Nobody else has access to your customers’ complaints, your sales records, your missed calls or your staff’s observations. That information is the one ingredient your competitor cannot copy.
The winners in the next few years won’t be the businesses with the best AI tools. They’ll be the ones who collect and use their own customer insight systematically, then use AI to act on it faster.
What to start collecting this week
- Why did your last ten customers choose you? Ask them, and write down their exact words.
- Why did your last five lost deals go elsewhere?
- Which questions do customers ask most on Zalo, Messenger and the phone?
- Which days, hours and products bring the most profit, not just the most revenue?
Give those answers to an AI and its advice changes completely. Suddenly it’s talking about your business, not an average one.
Where DMG fits
At DMG, we spend a lot of our first weeks with a client away from the screen: listening to sales calls, reading customer messages, studying the competitor next door. That ground-level knowledge becomes the foundation for every campaign, and it’s the part no AI can do for you.
Want to turn what your customers tell you into marketing that actually fits? Talk to DMG.
