ORHL: The Four Ways Your Brand Disappears Inside an AI Answer

Updated July 2026 · Ambika Sharma, Founder, Chief Strategist at Pulp Strategy Communications and Product Architect of NeuroRank^®

A brand can be missing from an AI answer in four distinct ways, and NeuroRank^® classifies them as ORHL: Omitted, Replaced, Hallucinated, and Zero Leads. Each one looks like the same problem from the outside, your brand is not winning the answer, but each has a different cause and a different fix. Naming which of the four you have is the difference between fixing the right thing and wasting a quarter on the wrong one. Here is what each class means and how to tell them apart.

Omitted: you are simply not there

Omitted means the model does not mention your brand at all for a prompt where it should. You are absent, and often a set of competitors fills the space. This is usually a source problem: the models are not reading enough about you in the places they trust, so you never enter the shortlist. The fix starts with the sources the models read for that prompt, not with your own page.

Replaced: a competitor holds your slot

Replaced means the model names a competitor where your brand belongs, for a query you should own. This is the most commercially painful class, because it is not neutral absence, it is a rival being recommended in your place. Replaced usually points to a competitor with stronger corroboration across the sources the model reads. The fix is to close that specific gap, not to shout louder on your own site.

Hallucinated: the model gets you wrong

Hallucinated means the model states something untrue about your brand, an outdated figure, a wrong feature, a product you no longer sell, or a claim you never made. This is the most urgent class, because the model is actively spreading a falsehood to buyers as if it were fact. The fix is to find the source the model is reading and correct it there. In one engagement, an outdated figure a model was citing back to a brand’s own customers was corrected within 40 days once the source was addressed.

Zero Leads: you appear, but it goes nowhere

Zero Leads means your brand shows up, but in a way that drives no consideration, a passing mention with no link, no recommendation, and no reason for the buyer to act. Presence without pull. This class is easy to miss, because a surface check says you appeared. The fix is to strengthen how you are described and cited, so a mention becomes a recommendation.

How to tell which one you have

You cannot fix what you have not classified. For each prompt your buyers actually ask, check whether you are absent, replaced, described wrongly, or present without pull, and note the source the model used. A brand often carries more than one class across its prompt set, which is why a single overall score hides the real work. NeuroRank runs this classification across ChatGPT, Gemini, Claude, and Perplexity, tags every gap by its ORHL class, and links it to the source the model read, so the fix is specific.