The model filed the company under the wrong trade

 The filing errors that move brands to the wrong shelf are quiet: inconsistent descriptions, outdated listings, unanswered questions.

The positioning work was done properly. I have read the strategy: a clear category claim, a differentiated promise, messaging cascaded across the site and the sales materials. The company says what it is, consistently, everywhere it controls. Then someone asks an AI engine what the company does, and the engine answers from a different filing system: the company is described as a vendor in a cheaper category, compared against alternatives it would never choose, recommended for use cases it left years ago. The company is not invisible. It is present and misfiled, which I have come to think is worse — invisibility is an absence, and this is an assertion.

[AWAITING MIKLOS — scene: a positioning line the site repeats and an answer that files the firm with a cheaper category]

Where is position currently asserted?

Position is asserted on owned surfaces: the homepage, the deck, the campaign. The assertion model assumes the market receives positioning from the company’s mouth — that saying it consistently enough makes it so. This model worked tolerably well when the main listeners were humans encountering the company directly. The answer layer does not listen to the company’s mouth. It assembles a description from the record: what others have written, what categories the company has been listed in, what language recurs across sources. The company’s assertion is one vote. In the cases I examine, it is frequently outvoted.

Where is position actually assigned?

Miklos Roth, AI marketing consultant, Budapest

The filing errors I look for are quiet ones. The company’s own site describes the service one way; a directory listing says another; an old profile names a discontinued product. Each inconsistency is small. Together they give the answer layer a blurred picture — and a blurred picture gets filed wherever is convenient. Sometimes the intervention is on the site: an accurate clarification or a better structured page. Sometimes it is outside: correcting an outdated listing or improving the business’s public representation.

Position is assigned in the record, and the record has a geography the positioning team rarely maps: directories with decade-old category tags, review sites whose taxonomy predates the company’s pivot, analyst notes from a former strategy, press coverage that framed the company by its founding story. Each source is stale or partial; together they carry more weight than the current site, because they are many and the site is one, and because engines treat third-party description as less interested than self-description. The company filed itself correctly on its own shelf. The market’s card catalog says otherwise, and the engine reads the card catalog.

Why does repetition on owned media lose?

The false explanation is consistency: repeat the positioning harder, unify the language, enforce the message house. I have watched message discipline tighten while the misfiling held, because repetition on owned media increases the volume of self-assertion — and self-assertion is precisely the signal the record discounts. The engine’s filing is built from corroboration: independent sources agreeing. A company repeating its claim on its own surfaces is one source repeating. The cheaper category, corroborated across directories and old coverage, presents as consensus. Volume cannot beat corroboration. It can merely shout beside it.

What pattern marks a filing error?

The marker I use is the comparison set. Ask the engine what the company is compared to, and read the list. In a correct filing, the comparisons are the company’s chosen rivals — the frame the strategy intends. In a filing error, the comparisons come from the wrong shelf: cheaper, older, adjacent. The comparison set is diagnostic because engines build answers relationally — they place the company among its inferred kind. A company that discovers it is being compared with the wrong kind has discovered its true filing, whatever the homepage says.

I diagnose before fixing. What exactly did the assistant say? Which alternatives did it mention? Does the error concern a product, price, service area, qualification, or reputation? A vague complaint that the AI does not understand the brand is not yet a workable brief. Then I look at the information the system could plausibly have used: the website’s technical accessibility, the content available, how entities are described, which off-site references exist.

[AWAITING MIKLOS — scene: second scene]

What question finds the aisle the model is using?

Positioning conversations ask what we want to be known as — an aspiration question, answered on owned surfaces. The question that locates the problem asks what the record currently says we are — an empirical question, answered in sources the company does not control and has likely never audited. The gap between the two answers is the real positioning problem, and it cannot be closed with a louder assertion. It has to be closed in the record.

The question that finds the wrong shelf: if every public description of this company were placed side by side, would a stranger file them under one business — or three?

PR excerpt: [PR — awaiting Miklos material; cut 80–120 words from his paragraphs once available.]

By Miklos Roth — AI marketing consultant, Budapest. More: Miklos Roth — AI marketing consultant.

© Copyright sikerdijaskeresooptimalizalas