Why ChatGPT, Perplexity & Claude Cite Different Businesses

If you've ever searched for your business across ChatGPT, Perplexity, and Claude, you've probably noticed something odd: you might appear in one, get ignored by another, and show up with completely wrong information in a third. That's not a glitch. Each AI platform decides which businesses to cite using a different set of signals — and understanding how ChatGPT, Perplexity, and Claude each make that decision is the fastest way to fix why your business keeps getting skipped.

They're Not All Doing the Same Thing

Most people assume AI systems work the same way search engines do — crawl the web, index pages, rank results. The reality is messier and more interesting. Some AI platforms retrieve live information at query time. Others rely entirely on what was baked into their training data months ago. Some weight structured data heavily. Others barely register it. The result is that a single business website can score completely differently across platforms — not because the site is inconsistent, but because each platform is asking a different question about it.

How Perplexity Decides Who to Cite

Perplexity is the most retrieval-forward of the major AI platforms. When someone asks it to recommend a local accountant or a SaaS tool, it actively crawls the web in real time, pulls sources, and synthesizes an answer from what it finds. That means recency matters. So does being findable — clear page structure, fast load times, and direct language that tells Perplexity exactly what your business does and where it operates.

• —Perplexity actively retrieves live web content at query time
• —It favors pages that are clearly structured and easy to parse
• —Business category, location, and services stated plainly on-page carry real weight
• —Third-party mentions and review sites (Yelp, G2, Trustpilot) feed its citations directly

How ChatGPT Decides Who to Cite

ChatGPT's behavior depends heavily on which version is being used and whether browsing is enabled. Base GPT-4o without browsing relies on training data — meaning a business that wasn't well-represented on the web before the training cutoff essentially doesn't exist to it. With browsing enabled, ChatGPT behaves closer to Perplexity, but it's still more conservative about citing specific businesses. It tends to recommend categories of businesses rather than specific names, unless the business has a clear, well-documented presence across multiple authoritative sources.

• —Without browsing: relies on training data, heavily favors established or widely-mentioned brands
• —With browsing: retrieves live content but applies more caution about specific recommendations
• —Schema markup and structured data help ChatGPT parse what your business actually does
• —Consistent NAP (Name, Address, Phone) data across directories reduces ambiguity

How Claude Decides Who to Cite

Claude — built by Anthropic — is training-data-dependent by default, with no live web retrieval in standard use. That changes things significantly. Claude's citations reflect what was prominent and credible on the web during its training window. It's also notably cautious about recommending specific businesses, especially local ones, because Anthropic has trained it to hedge on recommendations it can't verify. The counterintuitive insight here: having a Wikipedia entry, a well-structured Crunchbase profile, or meaningful press coverage matters more for Claude than almost any on-site optimization.

• —Claude pulls from training data — no live retrieval in standard mode
• —Authoritative off-site sources (Wikipedia, Crunchbase, industry publications) carry outsized weight
• —Claude is more likely to cite a business category than a specific company unless the brand is well-documented
• —Repeating consistent information across many trusted sources over time is the primary lever

The Counterintuitive Part

Here's what most people get wrong: they assume that ranking well in Google means AI systems will find and cite them. It helps, but it's not sufficient — and in some cases it doesn't help at all. Claude doesn't care about your current Google ranking. Perplexity cares about your page structure more than your domain authority. ChatGPT in training-data mode cares about how many times your business was mentioned across credible sources — years ago. You can have a page-one Google ranking and still be completely invisible to all three. The signals are genuinely different.

Why Your Score Varies Across Platforms

When Zygur scores a site, we evaluate it across the signals each platform actually uses — not a generic checklist. A local HVAC company might score well for Perplexity (clear location, service pages, review presence) but poorly for Claude (almost no off-site authority, no structured business data). An established SaaS tool might score well for ChatGPT training-data mode but poorly for Perplexity because the website is a mess of JavaScript that's hard to parse. The score varies because the platforms vary. Optimizing for all three requires knowing which gaps you're actually dealing with.

What Actually Moves the Needle Across All Three

• —State what your business does in plain language, in the first paragraph of your homepage
• —Make your business category, location, and services machine-readable — Schema markup isn't optional anymore
• —Build presence on third-party platforms AI systems treat as authoritative: Google Business Profile, industry directories, review sites
• —Get mentioned in external content — press, partner sites, industry roundups — not just on your own pages
• —Keep NAP data consistent across every platform it appears on
• —Use an llms.txt file to give AI crawlers a direct summary of what your business does

Frequently Asked Questions

Why does my business show up in Perplexity but not ChatGPT?

Perplexity retrieves live web content, so if your site is well-structured and current, it can find you. ChatGPT in its base mode relies on training data, which has a cutoff date. If your business wasn't well-documented on the web before that cutoff, you may simply not exist in ChatGPT's knowledge base — regardless of how good your site is today.

Does Google ranking affect AI visibility?

Indirectly, yes. High Google rankings often correlate with the factors AI systems also care about — credible backlinks, clear content, structured data. But they're not the same thing. A site can rank on page one in Google and score poorly for AI visibility because the page content is hard to parse, lacks structured data, or has no off-site authority signals that AI training pipelines picked up.

What is an llms.txt file and does it actually help?

An llms.txt file is a plain-text document placed at your site's root (yoursite.com/llms.txt) that gives AI crawlers a direct, structured summary of your business — what you do, who you serve, and what pages matter. It's a relatively new standard, but Perplexity and AI agents that crawl the web do use it. Think of it as a robots.txt, but designed to help AI understand your business rather than restrict crawlers.

How often do AI systems update their information about a business?

It depends entirely on the platform. Perplexity updates in real time — it's retrieving fresh content at the moment of the query. ChatGPT's training data has a fixed cutoff, updated only when OpenAI releases a new model. Claude is similar. This means for training-data-dependent platforms, building your off-site presence now is an investment in where you'll appear after the next training cycle — not an immediate fix.

Can I see exactly why my business scores differently across AI platforms?

Yes. Zygur's free score tool breaks down your AI visibility score by the specific signals each platform uses. Enter your URL and you'll get a 0-100 score in about 60 seconds, along with the specific gaps — missing structured data, weak off-site presence, unclear page content — that are pulling your score down on each platform.