# Why AI recommends only 4-7 vendors in your category > Google returns 10 results. AI usually returns 4-7 names. There's no page two. If you're not in the set, you're not in the evaluation. - Author: Max Wiesner - Published: 2026-08-14 - Canonical: https://knitknot.ai/blog/ai-recommends-only-4-7-vendors/ - Publisher: KnitKnot, the AI Competitive Positioning platform (https://knitknot.ai) --- ## How many vendors does AI actually name? Four to seven. When a buyer searches Google for "best compliance automation tools," they get ten results on page one, more on page two, and ads alongside everything. Even if they only evaluate three vendors, they're aware of ten. When a buyer asks ChatGPT the same question, the response names four to seven companies. That's it. No page two. No "see more results." No ads providing supplementary visibility. The AI synthesized the category, picked its answer, and moved on. We see this consistently in our benchmarks. Our prompt libraries are generated from real Google search data, with search volume attached to each prompt, so the category queries we test ("best [tool] for [use case]," "top [category] platforms") are the ones buyers actually type. The vendor set varies slightly by model, but the count is remarkably consistent: four to seven names. ## How concentrated are AI recommendations? Our benchmark answers concentrate recommendations among a small set of vendors. The names vary by model, but most category answers stop after four to seven. If a category has 30 vendors, most never appear in the answer. Compare this to Google, where the first page has 10 organic results plus paid placements. Position 8 on Google still gets impressions. Position "not in the response" in an AI answer gets nothing. AI search concentrates buyer attention at a rate search engines never did, and it does so without the buyer knowing alternatives exist. A Google user who doesn't click result #6 still sees the brand name. An AI user who gets back five names has no indication a sixth exists. ## Why 4-7? The number isn't arbitrary. It reflects how AI models construct category responses. **AI models produce short answers.** Listing 20 vendors would turn a recommendation into a directory. In our category benchmarks, four to seven names is the common range. **The AI needs enough source material to include a vendor.** Each company in the response gets a sentence or a paragraph. If the model doesn't have enough information to say something specific about your company, it leaves you out rather than including you with a vague mention. **Training data follows a power law.** Larger, more established companies have more web content, more reviews, more third-party mentions, more comparison pages. Smaller or newer companies may have excellent products but insufficient public content for the AI to cite. ## What determines who makes the cut? From our benchmark data and [source analysis](/blog/citations-are-ownership-claims), the companies that consistently appear in AI category responses share four traits. **They exist in the comparison content ecosystem.** They're mentioned in "best X tools" articles, G2 and Capterra reviews, Reddit discussions, and industry publications. The AI doesn't discover companies from their own product pages. It discovers them from the ecosystem of content that discusses the category. **They have structured, extractable information.** Not just a good website. Product schema, FAQ schema, plain-text pricing, feature pages with explicit capability lists. The AI needs machine-readable information to include you with specificity. **They're referenced by authoritative sources.** Domain authority matters for AI citations just as it does for SEO. A company cited by Search Engine Journal, TechCrunch, or Gartner has a citation advantage over a company cited only by its own blog. **They publish category-level content.** Not just content about themselves. "Best [category] tools for [use case]" pages, comparison guides, industry analysis. Companies that discuss the category position themselves as authorities on it, which makes the AI more likely to include them in category responses. ## What happens when you're not in the set? Three things, simultaneously. **The buyer doesn't evaluate you.** There's no page two to scroll to. No ad to catch their eye. If the AI didn't name you, you don't exist for that buyer. **The AI reinforces the exclusion.** The buyer's follow-up queries are about the companies that were named. "Tell me more about [vendor A]." "Compare [vendor A] vs [vendor B]." You aren't part of the follow-up conversation because you weren't part of the initial answer. **The gap compounds.** Named vendors accumulate more buyer interactions, more reviews, more content, more signals that reinforce their inclusion in future responses. Your absence from one query makes absence from the next slightly more likely. Over months, the disadvantage becomes structural. This is the winner-take-all dynamic that makes AI search fundamentally different from traditional search. On Google, every position gets some traffic. In AI, you're either in the answer or you're invisible, and invisibility compounds. ## How do you break into the AI answer set? There's no shortcut. Breaking in requires building the public information ecosystem that earns inclusion. **Get mentioned by third-party sources.** Guest posts, industry publications, analyst mentions. Every authoritative mention outside your own domain adds a signal the AI uses to determine category membership. **Engage in communities where AI pulls citations.** Reddit threads, Hacker News discussions, Stack Overflow answers. [Perplexity gets 46.7% of its citations from Reddit](/blog/why-ai-recommends-your-competitor). A detailed, useful post in a relevant subreddit can directly affect whether Perplexity includes you in category responses. **Publish category content, not just product content.** If you only write about your own product, the AI has no signal that you belong in the category conversation. A "Best [category] tools" page, even one that includes competitors, positions you as a category authority. **Build the technical infrastructure.** Submit your sitemap to Bing Webmaster Tools. Add Organization, Product, and FAQPage schema. Make your content [structured for extraction](/learn/content-formats-ai-models-love). The AI can't include information it can't find. The vendors already present in these answers keep accumulating reviews, comparisons, and follow-up questions. A missing vendor has to earn its way into a shortlist that buyers may never expand themselves.