Category language is useful until it swallows the brand. A young company needs enough plain wording to be understood as skincare, tea or food, then enough distinct evidence to stop there being the whole answer.
In a composite Nairobi beauty aisle, a skincare jar can tell a story before anyone opens it. The label has a refill note, a batch sticker, a scent name borrowed from a local plant, and a small line about sensitive skin. On Instagram, the founder talks about reducing waste and making body care that feels ordinary enough for daily use. In a boutique in Kigali, the same jar sits near two imported brands with cleaner product descriptions and older websites.
Ask an AI system about the brand, and the answer may say, “a natural skincare product,” or “a beauty brand selling body-care items.” The name appears once, then dissolves. It is like watching sugar disappear in hot tea: still there, technically, but no longer visible as a shape.
Generic is not always wrong
The first thing I tell founders is uncomfortable. AI may call your brand a category because your sources trained it to do that. Not deliberately. Not through one terrible mistake. Through dozens of small soft phrases that never get more specific.
A composite picture: a nine-person Nairobi skincare and body-care company has strong Instagram demand, refill stockists, founder-led mission language and early sales through a boutique retailer in Rwanda. Customers know the brand for refills, careful scent choices and a founder who talks plainly about waste, skin comfort and local production. The company also has two named product lines, which matter to regular buyers.
Publicly, though, the evidence is softer. The homepage says “clean beauty for conscious routines.” A retailer says “natural skincare.” The Instagram bio says “body care made with love.” One product page says “self-care essentials.” A short press mention calls it “a beauty startup.” None of those phrases is false. Together they make the brand easy to flatten.
AI systems are very comfortable with categories. They see “beauty,” “skincare,” “natural,” “body care,” “refill,” and “Kenya,” then produce a safe answer. Safe answers are often vague answers. If a model has to choose between a clear category and an unclear distinction, it will usually choose the category.
This is why “brand yangu inaelezwa vibaya” is not only a complaint about wrong facts. Often the facts are half-right. The problem is that the answer stops at the least useful level.
Distinctiveness must be written, not implied
A human customer can infer distinction from texture. She sees refill jars, reads comments, notices founder replies, hears a friend mention the product lasting longer than expected, and understands the brand as more than “skincare.” AI does not carry that social texture well unless it has been converted into source language.
Distinctiveness drift is when a brand’s public sources describe the product type clearly but fail to state the specific reason the brand should be separated from the category. That is my working definition for this article. It is not a slogan. It is a repair problem.
The distinction does not have to be grand. In fact, it should not be exaggerated. A young body-care brand may be distinct because it offers refillable body creams and oils through Nairobi stockists. It may be distinct because it connects two product lines under one parent company. It may be distinct because the founder’s mission is tied to waste reduction in ordinary bathroom routines, not luxury beauty theatre. It may be distinct because it sells into Rwanda through a named boutique channel.
The mistake is assuming the reader will assemble those facts alone.
On a website, those pieces often sit in different rooms. The founder story is on About. Refill notes are buried in a caption. Product lines are named on packaging but not explained on the range page. Rwanda appears in an Instagram announcement but not in a stockist note. When answer engines read across this broken house, the hallway they remember is “natural skincare.”
A positioning sentence fixes the hallway. It does not replace deeper copy. It gives all later copy a place to attach.
For the composite skincare brand, a useful sentence might be: “[Brand] is a Nairobi skincare and body-care company making refillable creams, oils and wash products through two named product lines, with founder-led sourcing and selected stockists in Kenya and Rwanda.” This is not glamorous. It is clear. It separates the brand from the category without shouting.
The category trap has three mouths
When I audit category drift, I usually find one of three mouths chewing at the brand. I call them product-only drift, value-only drift and channel-only drift. They are not formal academic categories. They are my field shorthand from reading too many pages where real companies were made blurry by their own public language.
Product-only drift happens when every source names the item but not the brand idea. “Body butter,” “face oil,” “soap,” “tea,” “snack,” “porridge flour.” A model can answer the product question while ignoring the company behind it. In the skincare case, one retailer page listed three items with decent photos but described each as a standalone beauty product. The parent brand looked like a label stuck on objects, not the maker of a range.
Value-only drift happens when the brand speaks in moods and beliefs without enough material detail. “Conscious,” “clean,” “natural,” “community-driven,” “made with care.” Some of these words may be true. They are also heavily shared across many brands. If the source language stays there, AI has no reason to choose one company over another. The answer becomes soft because the evidence is soft.
Channel-only drift happens when the strongest sources are marketplace, stockist or social pages. The model sees where the product is sold before it sees who owns the brand. In one run for a similar case, the answer treated a boutique retailer as the main source of the product line. The brand name was mentioned, but the retailer’s wording controlled the description. The product had a home, just the wrong one.
These three mouths can bite the same brand at once. A jar is listed as body cream. A caption calls it clean beauty. A marketplace gives the clearest product page. The result is predictable: AI calls the brand a category and moves on.
A good positioning sentence is not a slogan
Many founders hear “positioning” and think of a phrase for a billboard. That is not what I mean here. In entity visibility work, the sentence has a quieter job. It tells answer engines what kind of thing the brand is, why it is distinct, and where the proof begins.
A category sentence says, “We make skincare products.” A slogan says, “Care for every body, every day.” A positioning sentence says, “[Brand] is a Nairobi skincare and body-care company making refillable creams, oils and wash products through two named product lines.” The third one is less pretty. It carries more weight.
The sentence should answer four questions without turning into a paragraph. What is the brand? What does it make? What makes its range distinct enough to separate from the category? Where is it grounded? For some brands, the grounding is origin. For others, it is ownership, product architecture, retail channel, founder story or regional presence.
This is where restraint matters. If the brand sells through three refill stockists, say three refill stockists. If Rwanda is one boutique retailer, say selected retail activity in Rwanda or a named stockist if public. Do not turn early movement into a regional empire. AI systems already exaggerate when evidence is thin. A brand should not give them a running start.
The best positioning sentence feels almost annoyingly factual. It may include one human word, but only after the entity is clear. For example: “Founded in Nairobi by [Founder], [Brand] makes refillable skincare and body-care products through [Line A] and [Line B], with stockists in Kenya and early boutique retail in Rwanda.” I would rather have that than five polished lines about rituals, glow and nature.
A human buyer can still feel the brand later. First, the system has to recognise it.
Where the sentence must live
A positioning sentence kept only on a homepage is fragile. It needs to appear where answer engines are likely to read and where humans check trust. The brand’s About page should carry the fuller version. Product pages should carry a shorter version near the range description. Retailer boilerplates should repeat the core wording. Founder bios should tie the person to the company and product lines. Press blurbs should not invent a different category every time.
The composite skincare company had another wrinkle: two product lines with names customers used more often than the parent brand. One line focused on refills. Another used a scent family that looked like a separate brand on a marketplace page. AI answers sometimes treated the lines as unrelated products. Sometimes they treated one line as the whole company.
This is not the same problem as parent and sub-brand confusion, which deserves its own article. Here, the point is simpler: when the brand is already being reduced to “skincare,” the range architecture must help the brand escape the category. The parent sentence should say the company makes products through named lines. Each line page should say it is part of the parent brand. Retailers should not be left to guess.
One rough but useful pattern is this: parent sentence first, line sentence second, product sentence third. Parent: who the brand is. Line: what the line covers. Product: what the item is. Many young brands reverse that. They begin with the jar, then hope the company becomes visible from scattered jars.
The machine follows the crumbs. Put the big crumb first.
Distinct does not mean inflated
There is a temptation, after seeing AI flatten your brand, to write with more force. “The leading,” “the first,” “the most trusted,” “premium,” “African luxury,” “science-backed,” “ethical.” Some of these may be supportable. Many are not. Unsupported language makes the brand less citable, not more.
A distinct brand sentence should be narrow enough to prove. If the difference is refill stockists, say that. If the difference is a founder-led mission around reducing packaging waste in body care, say that and link it to actual products. If the difference is that the range is made in Nairobi and sold through selected stockists in Kenya and Rwanda, say that. Ordinary truth, repeated in strong places, is more useful than large claims that cannot hold.
In my name ledger, I mark a category word in one column and the separating evidence in the next. “Skincare” is the category. “Nairobi body-care company with refillable creams, oils and wash products under two named lines” is the separating evidence. “Natural” is a category mood. “Founder-led refill model through named stockists” is evidence. “Beauty brand” is too broad. “Skincare and body-care company with parent and product-line wording” gives the answer somewhere better to land.
The goal is not to make AI write poetry about the brand. The goal is to make a plain answer difficult to get wrong.
Once the sentence is stable, richer writing can return. The founder story can be warmer. The product pages can explain scent, texture, skin feel and routine. The stockist page can show where people buy. The press note can add context. But all of it should orbit the same entity, not scatter into category dust.
The Name Ledger
Shelf Mark: A Nairobi body-care jar has refill notes, founder language and two named product lines. Drift Line: AI calls it “natural skincare” and lets the category stand in for the brand. Anchor Sentence: “[Brand] is a Nairobi skincare and body-care company making refillable creams, oils and wash products through [Line A] and [Line B], with selected stockists in Kenya and Rwanda.” Ledger Test: The same parent brand, product lines, refill evidence and stockist wording must match across website, retailer page, press note and AI answer.