AI search · pain-led

Why ChatGPT doesn't recommend your business yet.

If your business is invisible to ChatGPT, Claude and Gemini, it's almost always one of five fixable reasons. Here's what's actually going on — and what to do about it.

You've spent money on SEO. Your site looks fine. Google more or less knows you exist. But when a real customer asks ChatGPT, Claude or Gemini for businesses in your category, you don't get mentioned. Your competitor does. That's the painful part: they're not always better than you. They're just structured in a way the AI can quote.

Here are the five reasons AI engines aren't recommending your business yet, in order of how common they are. If any of these feel familiar, you're not alone — most websites we audit fail on three or four of them.

1. The AI literally can't see your content properly

Large language models read websites differently from humans. Where you see a clean homepage with a hero image and a navigation bar, an AI sees a stream of text and structure tags. If your site relies heavily on JavaScript to render content (typical of Wix, Webflow, Framer or custom React builds), there's a real chance the AI's crawler is hitting your site and seeing... an empty shell.

The fix: server-side render the parts of your site that describe what you do. Make sure your <title>, <meta description>, headings and primary content are in the initial HTML response, not injected by JavaScript after page load.

Quick test: open site:yoursite.com in Google. If you see your individual pages with their full titles and meta descriptions, you're fine. If you see "JavaScript required" or generic placeholder text, this is your number-one fix.

2. You have no structured data — so the AI has to guess

Schema.org is the universal vocabulary AI engines use to understand websites. It's a set of JSON-LD snippets that tell the AI explicitly: this is a business, located here, that sells these things, with this rating, run by these people. Without it, the AI has to infer all of that from unstructured paragraphs — and inference is where competitors quietly overtake you.

The five schema types that matter most for AI citation, in order:

Adding these properly typically takes a few hours of work. The impact on AI citation rate is one of the strongest single-fix interventions we see.

3. Your domain has no authority signals AI engines trust

AI engines err heavily on the side of recommending sources they perceive as credible. Credibility comes from things that exist off your website: backlinks from established domains, mentions on industry publications, presence in directories, consistent profiles across Google Business, LinkedIn, Wikipedia, niche review sites.

If your domain is six months old with no backlinks and no off-site mentions, it doesn't matter how good your on-site content is — the AI will skip you in favour of the slightly worse competitor who has a 12-year-old domain and a Forbes mention from 2018.

This is the hardest part to fix because it can't be done in a weekend. It needs months of relationship-building, content placement, podcast appearances, directory submissions and PR. It's also where most agencies quietly stop doing actual work and just bill you anyway.

The honest version On-site fixes can take you from "invisible" to "occasionally cited" within weeks. Going from "occasionally cited" to "the AI's default answer for your category" is an off-site authority project that takes months. There's no shortcut. Anyone selling you one is selling you nothing.

4. Your content doesn't actually answer the question the AI is being asked

When someone asks ChatGPT "what's the best AI visibility tool for small businesses?" the model isn't searching for "AI visibility tool". It's looking for content that directly addresses that specific phrasing — the kind of buyer-intent, comparison-shaped question a customer would actually type.

Most business websites are written in marketing voice. Hero headlines, benefit-led bullet points, calls to action. That's fine for converting a visitor who's already on the page, but it's terrible for getting cited by AI, because none of it matches the conversational, question-shaped pattern of a real customer query.

The fix: add genuine question-and-answer content. FAQ sections. Blog posts with question-shaped headlines. Comparison pages. Each one written to answer a real query, not to sell.

5. You haven't told AI engines you exist

Some practical mechanics that quietly stop sites from being indexed:

So what should you actually do?

In order of impact-to-effort:

  1. Run a real audit on your site to see which of the five reasons above are affecting you. Top Level App does this and shows you exactly which signals are missing, ranked by what to fix first. Most sites fail on three or four — fixing the top one or two is usually enough to start showing up.
  2. Add Organization, FAQPage and (where relevant) LocalBusiness or Product schema. This is a weekend of work.
  3. Make sure your robots.txt explicitly welcomes the AI bots and you have a sitemap.xml.
  4. Write three to five Q&A-shaped pages targeting the actual questions your buyers ask AI engines.
  5. Start the slow off-site authority work: directory submissions, niche publication outreach, podcast guesting. Allocate a small budget for this and stay patient.

Want to know exactly which of the five it is for your site?

Run your URL through Top Level App and you'll get a scored breakdown of which signals are missing, in plain English, ordered by ROI.

See plans — from £49/mo

Final note

None of this is magic. AI search optimisation is just rigorous, honest work applied to the same fundamentals that have always mattered: clean structure, real authority, useful content. The tools change. The principle doesn't. Most businesses we audit have one or two genuinely broken signals — fix those, and your visibility shifts within weeks.