If you ask ChatGPT or Perplexity for the best option in your category and your brand does not come up, it is almost never bad luck. It is one of seven specific, fixable reasons. Your content might be unreadable to AI crawlers, unstructured in a way that makes it impossible to quote, absent from the third-party sources models actually trust, or attached to a brand the model cannot clearly identify. This article walks through the real reasons, in the order they usually matter, so you can find yours.
Start here: confirm you actually have a problem
Before diagnosing, measure. Open ChatGPT, Perplexity, Gemini, and Claude, and ask each the question your best customer would ask before buying, the real one, like "best inventory software for a small business," not "is [your brand] good."
Note two things: whether you appear at all, and if you do, whether you are named first or buried in a list. This is the difference between presence and prominence, and it matters, because being mentioned sixth is functionally invisible. Do this across all four platforms, because they disagree with each other more than people expect. Now, the reasons.
Reason 1: AI literally cannot read your site
The most basic and most common cause. If your pages are not indexed, or an AI model cannot crawl and render them, you do not exist as far as that model is concerned. New sites, JavaScript-heavy sites that do not render cleanly, and sites with technical errors all fall into this category.
There is a specific version of this that catches people out: blocking the wrong crawlers. "AI crawlers" are two different groups. Training crawlers (GPTBot, ClaudeBot) build model datasets, and blocking them does nothing to your citations. Retrieval crawlers (OAI-SearchBot, ChatGPT-User, PerplexityBot, Claude-SearchBot) fetch pages for live answers, and blocking those removes you from AI results entirely. Plenty of sites block the second group by accident. We cover the full distinction in our guide to AI visibility.
Fix: get indexed, make sure your pages render server-side, and check your robots.txt allows the retrieval crawlers.
Reason 2: You rank on Google, but ranking is not citation
This is the one that surprises marketers most. You can hold position one for your main keyword and still be absent from the AI answer for the same question. Ranking and being cited have come apart: Ahrefs found only 38% of AI Overview citations now come from pages in Google's top ten, down from 76% just eight months earlier.
So "we rank well" is not evidence that you will show up in AI answers. They are related but increasingly separate games, which is the whole reason AI search differs from traditional search.
Fix: stop treating rankings as proof of AI visibility. Measure AI visibility directly.
Reason 3: Your content cannot be quoted
Even when a model can read your page, it may not be able to use it. AI answers are built by lifting clean, self-contained passages. If your content buries the answer in long prose, hedges instead of stating things directly, or never actually answers the question a buyer asked, there is nothing for the model to extract.
This is the answer engine optimization problem, and it is more common than it looks. In our own audits, brands' blog content consistently scored lowest on extractability, not because the information was missing, but because it was not in a shape a machine could quote.
Fix: answer the question directly in the first sentence, use clear question-style headings, and make each section stand on its own.
Reason 4: You are absent from the sources AI actually trusts
Here is the uncomfortable one. Most of what AI says about your brand does not come from your website. When we traced where AI pulled its citations for one brand across 113 prompts, its own site accounted for only about 6% of the citations. The other 94% came from reviews, comparison articles, community threads, and video. The single most-cited source was not the brand or any competitor. It was YouTube.
If you are not present in those third-party places, models have nothing to build a recommendation from, no matter how good your own site is. This is generative engine optimization: the off-site authority that makes a model confident naming you. Brand mentions across the web correlate with AI visibility about three times more strongly than backlinks.
Fix: earn genuine presence where your category is discussed and cited, from review sites and editorial coverage to relevant communities and video.
Reason 5: The model cannot tell who you are
A model can only recommend a brand it can cleanly identify. If your name is ambiguous, if the web describes you inconsistently, or if the model holds stale or wrong information about you, it has no clear entity to attach a recommendation to. Sometimes the model even confuses you with a different company that shares your name.
This is the entity problem, the layer LLMO points at. It is easy to miss because your content can be perfect while the model still cannot place you.
Fix: make your brand unambiguous and consistent everywhere: the same name, description, and category across your site, profiles, and every third-party source.
Reason 6: You appear, but you are buried
Sometimes you are not absent, you are just not prominent. You show up in long lists, or on easy discovery questions, but vanish on the high-intent comparison and workflow questions that actually precede a purchase.
We saw exactly this in our first AI Lab study: a brand that led its category on overall share of voice was still missing from 35% of the questions its buyers ask, specifically the comparison and how-to prompts, because competitors had built content for those and it had not. A category-leading average hid a serious gap.
Fix: track prominence on your highest-intent questions, not just an overall visibility score, and build content for the specific decisions where you disappear.
Reason 7: You are too new, or too quiet
Models learn who matters from how often and how consistently credible sources talk about a brand. A young brand, or one that few people discuss, simply has not accumulated enough signal yet. This is not a bug you can patch in a week; it is authority you build over months.
Fix: there is no shortcut. Consistent content, genuine mentions, and real coverage compound over time. The brands cited today are largely the ones that started building signal early.
How to figure out which reason is yours
Work down the list in order. Can AI read your site at all? Are you indexed and are the retrieval crawlers allowed? Is your content actually quotable? Are you present in the third-party sources models cite? Can the model identify you correctly? Are you present but buried? Are you simply too new?
Most brands have two or three of these at once, and the mix is specific to you. The fastest way to see it clearly is a structured audit across every platform, which is exactly what a proper AI visibility diagnosis does.
Frequently asked questions
Not sure which of these is costing you? Get a free AI Visibility Audit and see exactly where your brand is missing, and who is getting cited instead.
Sources
- Ahrefs, "Update: 38% of AI Overview Citations Pull From The Top 10" (863K SERPs, 4M AI Overview URLs, March 2026) — https://ahrefs.com/blog/ai-overview-citations-top-10/
- Ahrefs, "76% of AI Overview Citations Pull From the Top 10" (July 2025 baseline) — https://ahrefs.com/blog/search-rankings-ai-citations/
- Ahrefs, "An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)" (brand mentions vs backlinks) — https://ahrefs.com/blog/ai-overview-brand-correlation/
- OpenAI, "Overview of OpenAI Crawlers" (OAI-SearchBot vs GPTBot) — https://developers.openai.com/api/docs/bots
- Anthropic, crawler documentation (ClaudeBot vs Claude-SearchBot / Claude-User) — https://support.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler
- Slate AI Lab, AI Search Visibility Audit (113 prompts, 6 platforms; ~6% of citations from owned site; 35% of prompts with critical gaps), proprietary Slate AI Lab research, methodology available on request, April 2026.