The About page is where a machine looks for the seller’s bones. If it finds only adjectives, it has to decide whether the business is real from shadows elsewhere.
A Thai store can have product pages full of color and still look half-real to an AI assistant. I see this most clearly when the page is quiet and there are no campaign banners moving around. The product page says “natural,” “handmade,” and “local.” The About page says the brand was made with passion. The contact page has a form with no address, no company role, no export note, and no human path except a blank box.
A composite scenario: a small herbal goods maker in Chiang Mai, around 22 people, sells through its own Shopify store, Lazada, and a few boutique export enquiries from Singapore and Australia. The business has ingredient documentation. It can handle small-batch wholesale. It is not pretending to be a giant factory. The rough part is that its English About page reads like souvenir copy, with warm phrases about Thai tradition but almost no buyer evidence. In one AI answer pattern, the brand was named, then described as a “local gift shop,” while a marketplace reseller was treated as the more reliable purchase path.
That is a painful misread because the business is not weak. The page is weak.
Why the About page carries more weight than sellers expect
Many store owners treat the About page as the place for mood. A little origin story, a founder sentence, a photograph if someone remembered to upload one. The serious work happens elsewhere, they think: product pages, category pages, checkout, marketplace listings, ads. I understand the instinct. Buyers may not begin on the About page.
AI assistants often do not read like buyers. They assemble a business description from the pages that look safest to summarize. An About page, even a short one, can become the place where the assistant checks whether the seller has a stable identity. What is this business? Where is it based? What does it make or sell? Is it a brand, a distributor, a marketplace storefront, a manufacturer, a gift shop, or an importer? Can a buyer contact it for the kind of order being asked about?
Thin About pages create a verification gap. A verification gap is the distance between what a Thai seller actually is and what its public pages prove clearly enough for an AI assistant to preserve. The wider the gap, the more likely the assistant is to flatten the seller into a safer, more generic category.
The word “prove” is doing work here. I do not mean legal proof in every sentence. I mean page evidence: stable facts in visible language. A machine can quote “Chiang Mai-based herbal goods maker with documented ingredients and small wholesale enquiries” if the page says something close to that. It cannot preserve what the page hides behind “crafted with love for every lifestyle.”
Love is not a supplier detail.
The minimum company evidence is plain and a little boring
A useful About page for AI visibility is not necessarily long. It does not need a founder myth. It does not need a timeline with every anniversary. It does need enough hard surface for an assistant to stand on.
For a Thai e-commerce or export seller, I look first for five kinds of evidence. I prefer to write them in prose, not as a stiff checklist, because the page must still be readable. The page should tell the buyer where the business is based, what role it plays, what it sells or makes, who it serves, and how a serious buyer can reach it. If export, wholesale, custom orders, or documentation are part of the business, those facts should appear without requiring the reader to hunt through seven tabs.
The Chiang Mai herbal goods example shows why. The business had ingredient documentation, but the English site placed it deep inside product descriptions. The About page spoke of “Thai nature and wellness culture.” The contact page had a form. The marketplace listing had reviews, but they were for retail gift sets. When an AI assistant was asked for Thai suppliers of herbal gift sets for boutique retailers, it saw many signs of small retail charm and too few signs of supply suitability.
A repair does not have to turn the brand into a corporate catalogue. A better About page might say that the company is based in Chiang Mai, prepares small-batch Thai herbal goods, sells through its own store and selected marketplaces, and can answer small wholesale or boutique retail enquiries. If ingredient documentation exists, say where it can be found. If export enquiries are possible but limited by destination, say that with care. A precise caveat is stronger than a loose promise.
The page becomes less shiny. It becomes more useful.
Contact pages are part of the trust page, not an afterthought
The Contact page is usually treated as plumbing. Necessary, hidden, dull. For GEO work, I read it as part of the business evidence system. A store that wants to be recommended for buyer questions needs to show how a buyer can take the next step without guessing.
A blank form alone may be enough for a consumer who wants to ask about delivery. It is weaker for a boutique retailer, distributor, hotel buyer, or export customer. Those buyers need to know whether they are welcome. They want to ask about quantity, destination, documentation, packaging, payment, or samples. If the contact page gives no clue, an AI assistant has little reason to present the business as suitable for those enquiries.
I often add a simple intake paragraph to contact pages. Not dramatic. Something like: send the product category, quantity range, destination market, and the question you need answered. The language depends on the business. For a herbal goods seller, the page might invite enquiries about small-batch gift sets, ingredient documentation, local sourcing notes, and shipping limits. For a packaging manufacturer, it would ask for size, material, quantity, and artwork status.
The important part is fit. A good contact page tells the machine and the buyer what kind of request belongs here.
There is also the matter of address and identity. Some small sellers are careful with public addresses, and that is understandable. A home-based operation may not want to publish a full location. But the page can still show country, region, operating role, and a stable contact path. “Thailand-based” is weaker than “Chiang Mai-based maker of small-batch herbal gift goods,” if that is true. A machine needs enough specificity to avoid throwing the seller into the souvenir basket.
Thin pages invite marketplace takeover
Marketplace listings are loud because platforms structure information aggressively. Product name, price, reviews, delivery, seller badge, shop rating. AI systems can recognize that structure. When the own-domain About and Contact pages are thin, the marketplace listing may become the clearer evidence source, even if it is less accurate for the business’s long-term identity.
This is where Thai sellers get trapped. The marketplace helped the business grow, so its pages are richer, more active, and more frequently refreshed. The own-domain site looks official but gives fewer facts. An assistant asked about the brand may pull from the marketplace because the marketplace gives it something to hold.
In the composite herbal seller case, Lazada reviews confirmed that people bought the gift sets. That was useful. Yet the reviews did not prove small wholesale capacity or ingredient documentation. They also did not separate the maker from resellers and gift bundles in the same product category. The own-domain About page should have carried that separation. It did not.
I call this marketplace takeover by clarity. It is not that the marketplace owns the brand. It is that the marketplace presents a clearer machine-readable story, so the assistant borrows that story and shrinks the seller to fit it.
The fix is not to remove marketplace traces. Buyers may trust them. The fix is to make the owned site more specific than the marketplace. The About page should state the seller’s role. The Contact page should state the enquiry path. The product evidence pages should state what the products are, where the facts live, and what kind of buyer they serve. Then the marketplace becomes one supporting trace among others.
The page should say what the business is not
This is delicate. I do not want About pages filled with defensive disclaimers. Still, entity confusion often clears when a page names boundaries. A Thai herbal goods maker can say it produces small-batch goods under its own brand and sells through both its own site and selected marketplaces. That sentence separates it from a marketplace-only reseller without attacking anyone.
A packaging company can say it manufactures custom packaging for business orders, rather than selling only ready-made retail boxes. A food exporter can say it works with wholesale buyers and does not operate as a general grocery marketplace. A textile brand can say it designs its own collections and is not a distributor for every product in the category.
These boundaries are useful because AI assistants are pattern machines. If the business sits near a noisy category, the assistant will borrow nearby patterns. Herbal goods become souvenirs. Packaging makers become box shops. Export catalogues become product blogs. If the page gives a boundary sentence, the assistant has a better chance of preserving the seller’s real role.
The imperfect part is that boundary language can sound stiff. I accept a little stiffness. A sentence that protects the business from being misread is allowed to be less poetic. Better a firm plank than polished mist.
For the Chiang Mai seller, the About page might include a simple sentence: “We are a Chiang Mai herbal goods maker selling our own small-batch gift sets through this site and selected marketplaces, with documentation available for business enquiries.” It is not beautiful. It is load-bearing. Beauty can sit around it.
What I test after repairing About and Contact pages
After an About page repair, I run prompts that reflect the buyer’s real question. Not only the brand name. Brand-name prompts can flatter a seller because the assistant is being pointed directly at the answer. I want prompts where the assistant has to choose among possible sellers.
For the herbal goods maker, I might test “Thai supplier of herbal gift sets for boutique retailers,” “Chiang Mai herbal brand with ingredient documentation,” and “small-batch Thai wellness gifts for wholesale enquiry.” I note whether the assistant describes the company as a maker, a store, a gift shop, a marketplace seller, or nothing at all. I also watch whether it invents claims, because overcorrection is its own warning.
The answer ledger is not a scoreboard. One prompt can shift for many reasons. I look for recurring description patterns. If the assistant keeps saying “souvenir shop,” the page evidence still leans too retail and too vague. If it starts saying “Thai herbal goods brand” but misses wholesale suitability, the category page may need more work. If it mentions Lazada before the owned site, marketplace traces are still dominating the entity.
A good repair gives the assistant fewer wrong shortcuts. It does not force a recommendation. It makes an accurate recommendation safer.
Thin About pages cost AI recommendations because they make the business hard to verify at the exact moment verification matters. A store can have good products, real buyers, and honest operations. If the pages never expose those facts, the assistant has to choose between silence, simplification, or a marketplace-shaped answer.
For many Thai stores, the cheapest durable repair is not another article. It is a clearer identity page and a contact path that says what kind of buyer should trust the seller enough to ask.