One of the most common questions we get at FlinnSchema is some version of this: "How do I actually get ChatGPT to mention my business?" It is a fair question and a useful one to ask, because most businesses are completely invisible to ChatGPT and the other major AI engines, and most do not know why. This guide answers the question in detail, covering the technical foundations (structured data, llms.txt, crawler access), the content side (quality, conversational writing, E-E-A-T signals), and the priority order that actually moves the needle.
If you want the broader context first, our overview on what AI visibility is and why it matters sets the stage. The companion piece on whether customers actually use ChatGPT to find businesses covers why this matters commercially. This post focuses specifically on the how.
Why Most Businesses Are Invisible to ChatGPT
Before we get to the answer, it helps to understand the problem. When you ask ChatGPT something like "Who is the best recruitment agency in Kent for tech roles?" or "Which UK jewellery brands specialise in sustainable materials?", the model does not pull from a fixed directory. It sends real-time web search queries, retrieves the top results, extracts entity information from each page, and synthesises an answer mentioning specific businesses. The whole pipeline takes a few seconds and produces a written response with named recommendations.
Most businesses fail this pipeline at the very first step. Their pages might be technically indexable by Google but invisible to ChatGPT's specific crawler (GPTBot). Their content might rank well on Google but lack the structured data that lets ChatGPT extract clear facts. Their reviews might exist on third-party platforms but not be cross-referenced from their own site. The cumulative effect is that ChatGPT runs its retrieval, finds nothing useful from the business's site, and either skips them entirely or describes them generically without the brand-level recommendation that drives commercial outcomes.
We have audited hundreds of businesses on FlinnSchema, and the pattern is consistent. Businesses with strong traditional SEO often score below 30 percent on AI visibility because they have optimised for Google's signals rather than ChatGPT's. Closing that gap is what this guide is about. Our breakdown of how AI search engines decide which businesses to recommend covers the retrieval mechanics in more detail.
Structured Data: The Foundation Everything Else Builds On
Schema markup is the single most influential factor in whether ChatGPT mentions your business. The reason is mechanical: ChatGPT reads your pages in milliseconds during retrieval, and it needs to extract clear facts about your business (name, what you do, where you are, who reviews you, what you charge) before deciding whether to cite you. JSON-LD schema gives it that information in a machine-readable format. Without it, the AI has to guess from unstructured text, and guesses are unreliable.
For most businesses, the schema mix that matters is:
- Organisation on the homepage to identify your brand as a specific entity with contact details, social profiles, and founding information
- LocalBusiness or Product depending on your business model (service businesses use the LocalBusiness subtype that matches your category; e-commerce uses Product on every product page)
- Review and AggregateRating nested inside Organisation, LocalBusiness, or Product to surface verified customer feedback
- FAQPage on key landing pages, structured with the actual questions real customers ask in natural language
- BreadcrumbList on non-homepage URLs to give ChatGPT a clear picture of how your site is organised
Each schema type has required and recommended fields. Implementing them partially can do more harm than not implementing them at all because invalid schema makes ChatGPT skip your entire JSON-LD block. Our deep dive on which schemas your specific business needs walks through the decision by business type. For the technical reference, the Schema.org documentation is the canonical source, and our complete schema markup guide for 2026 shows the exact JSON-LD format for each type.
One client we worked with, a Shopify jewellery store, started with two basic schema types and an AI visibility score of 31. After we implemented eight complete schema types (Product, AggregateRating, Review, Offer, FAQPage, BreadcrumbList, WebSite, and Organisation), their score rose to 84 and ChatGPT began citing them by name for queries about sustainable jewellery brands. Their Google search impressions also rose 155 percent during the same period because complete schema unlocks rich snippets in traditional search as well. More before-and-after detail is on the FlinnSchema results page.
llms.txt: The AI-Specific Summary File
llms.txt is a plain-text markdown file hosted at the root of your website (yoursite.com/llms.txt) that gives AI engines a human-curated description of your business and a structured list of your most important pages. It emerged as a proposal aimed specifically at giving large language models a reliable entry point for understanding any website, and the spec is published at llmstxt.org.
The file is short, simple, and high-leverage. A good llms.txt includes a one-line description of your business, a paragraph of context, and grouped sections linking to your most important pages with brief descriptions of each. ChatGPT, Perplexity, Gemini, and Grok all retrieve and parse llms.txt when they find one. The impact varies by engine but the consistent benefit is that it gives the AI a clean source to draw from when introducing your business, rather than forcing it to construct a description from your homepage marketing copy.
One client (an artisan kitchenware brand we audited) saw ChatGPT's description of their business shift from a generic "UK kitchenware retailer" to an accurate "UK-based artisan kitchenware brand focused on hand-forged knives and traditional cookware made by independent makers" within weeks of adding llms.txt. The accuracy gain came entirely from giving the AI a clean, branded description to use.
For the full anatomy of the file, our deep dive on what llms.txt is and whether you need it covers the implementation. For the technical walkthrough, see our LLMs.txt: the new file every website needs in 2026 post.
Content Quality and Conversational Writing
Schema and llms.txt give ChatGPT the structural foundation. Your actual page content determines whether ChatGPT trusts your business enough to recommend it. The pattern we see across hundreds of audits is that businesses with the highest citation rates have content that does three things differently from average:
- Directly answers questions in the first paragraph. If a customer's likely query is "Do you offer mobile dog grooming in Maidstone?", a page that opens with "We offer mobile dog grooming across Maidstone, Tunbridge Wells, and the surrounding Kent area" tells ChatGPT exactly what to extract. A page that opens with marketing copy ("Award-winning pet care, trusted by thousands") tells ChatGPT nothing useful.
- Uses concrete, specific language rather than generic superlatives. "We specialise in elderly care planning including Lasting Power of Attorney, deputyship applications, and care home funding advice" beats "We provide expert legal services for families navigating life's challenges." The first is extractable as facts. The second is filler.
- Organises content under descriptive headings. ChatGPT's retrieval pipeline parses headings to understand topic structure. Headings like "What we cover" or "Our process" tell the AI very little. Headings like "Family law services in Kent" or "How long does an LPA application take?" map directly to the queries real customers ask.
This style of writing is sometimes called conversational content, and it is one of our 26 measured factors with a 1.5x impact multiplier. Pages written this way score consistently higher in AI visibility tests because they look more like the source material AI engines were trained to extract information from.
The connection to traditional SEO is real but limited. Some practices overlap (clear headings, helpful content) but the dominant style of SEO-optimised copy from the last decade (keyword density, vague brand-friendly language) often hurts AI visibility because it gives the AI nothing concrete to cite. Our piece on how AI visibility is different from SEO covers the broader divergence.
E-E-A-T Signals AI Engines Actually Use
E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) is Google's framework for evaluating content credibility. ChatGPT and the other major AI engines have their own version of this evaluation, and the signals overlap heavily with Google's. The difference is that AI engines lean harder on third-party verifiable signals because they cannot easily evaluate your own marketing claims.
The signals that move AI citation rates the most are:
- Third-party reviews on multiple platforms. A 4.7-star rating on Trustpilot with 200 reviews, a Google Business Profile with 80 reviews, and presence on industry-specific platforms (Bark, Checkatrade, Yell for trades; LegalChoices for solicitors; AccountingWEB for accountants) collectively form a credibility signal that AI engines treat as high-trust because they can independently verify it.
- Citations and mentions on third-party sites. If trusted publications, blogs, or forums (Reddit especially) mention your business, ChatGPT treats those as independent corroboration of what your own site says. This is one of the most underrated AI visibility levers for service businesses because organic mentions on niche subreddits can outweigh paid PR placements.
- Author and entity attribution. Content attributed to named individuals with clear credentials (Person schema with sameAs links to LinkedIn, professional bodies, or accreditation databases) scores higher than anonymous content. For consultancies and professional services, this is one of the biggest gaps we find on audits.
- Consistent NAP (name, address, phone) details across the web. If your business is listed with different addresses or phone numbers across directories, AI engines penalise the inconsistency because it creates uncertainty about which entity is canonical.
The Princeton GEO benchmark paper, available at arxiv.org/abs/2311.09735, tested specific content interventions including citation density and statistical specificity, and found that both significantly improve generative engine citation rates. This is consistent with what we see in practice across the businesses we audit.
Crawler Access: The Step Most People Forget
None of the above matters if AI engines cannot read your pages in the first place. The crawlers that retrieve content for ChatGPT, Perplexity, Gemini, and Grok have specific user agent names that your robots.txt and any hosting-level bot filtering need to allow.
The major AI crawlers to allow are:
- GPTBot (OpenAI)
- ClaudeBot (Anthropic)
- PerplexityBot (Perplexity)
- GoogleOther (used by Gemini for grounding)
- Bytespider (ByteDance)
Many hosting providers and security plugins block non-standard user agents by default. WordPress sites with Wordfence or similar plugins often block AI crawlers without the owner realising. Shopify sites generally allow them but some custom themes restrict bot access. The fix is to check your robots.txt at yoursite.com/robots.txt and look for any Disallow rules targeting these agents, then ensure your hosting layer is not adding additional blocks.
This single step has fixed AI visibility scores by 15 to 20 points in some of our audits. A recruitment agency client we worked with had GPTBot and ClaudeBot blocked at the hosting level (a leftover from a security audit that had blanket-blocked unfamiliar agents). After we unblocked them, their citation rate jumped within two weeks because the engines could finally retrieve their content.
Case Study: From Invisible to Mentioned
The clearest example we have of moving from invisible to consistently mentioned is a recruitment agency in Kent specialising in tech and engineering placements. When we first audited them, their AI visibility score was 18 out of 100. We sent twenty prompts about Kent and UK tech recruitment to ChatGPT, Perplexity, Gemini, and Grok. Across forty test queries, they were mentioned zero times. They had been a successful agency for over a decade, with strong Google rankings and good reviews, but every AI engine treated them as if they did not exist.
The work we did over eight weeks was unglamorous and methodical. We implemented complete Organisation schema, LocalBusiness with service area, Service blocks for each placement category, Review schema, and FAQPage on key landing pages. We added an llms.txt file describing the agency and linking to their main service pages. We opened up AI crawler access in their robots.txt. We restructured their key pages to answer common candidate and client questions directly. We consolidated their review signals across Google Business Profile, Trustpilot, and industry-specific platforms.
By the end of the eight weeks, their AI visibility score had risen to 62. In follow-up LLM testing, they appeared in twenty-three out of forty queries across the four engines. ChatGPT cited them by name in answers about Kent tech recruitment. Perplexity included them as a source. Gemini referenced their site in two answers. Grok mentioned them in two answers about regional engineering recruitment. Their Google rankings during this period changed by less than one position on tracked keywords. The two channels moved independently because they use different signals. More case study breakdowns are on the FlinnSchema results page.
The Honest Priority Order
If you take only one thing from this guide, make it the priority order. The highest-impact moves for getting ChatGPT to mention your business, ranked by what we see actually shift citation rates, are:
- Verify AI crawler access. Check your robots.txt and any hosting-level bot filtering. This is the cheapest and most-skipped step. Blocked crawlers mean nothing else matters.
- Implement complete Organisation and LocalBusiness or Product schema. Make these full and validated, not partial. Use the Google structured data reference for required fields, then test with Google's Rich Results Test.
- Add Review and AggregateRating schema linked to verifiable third-party platforms. Trustpilot or Google Business Profile at minimum, ideally both, with consistent NAP details.
- Restructure your highest-traffic pages to answer customer questions directly. Replace marketing copy with specific factual descriptions. Use descriptive headings that match real customer queries.
- Add llms.txt with a clear description, your most important pages grouped by category, and brief descriptions of each link.
- Add FAQPage schema on key landing pages with the actual questions customers ask, in their own phrasing.
- Build third-party citations by participating genuinely in relevant Reddit communities, industry forums, and getting listed on niche directories that AI engines treat as authoritative.
This order roughly matches our 26-factor weighting. The first three usually account for the largest jump in citation rate, often moving businesses from 0 out of 20 mentions to 8 to 12 out of 20 within four to eight weeks. The remaining items provide compound gains over the following months. Our guide on how to increase your AI visibility score covers the same priority order with more implementation detail.
Getting Started
The fastest practical first step is to run our free 26-factor audit. It takes about 60 seconds, requires no credit card, and produces a precise AI visibility score plus a list of the specific gaps holding you back. From there, you either implement the fixes yourself (the priority order above is a usable roadmap) or take advantage of our Premium plan which covers ongoing monitoring, daily LLM testing across ChatGPT, Perplexity, Gemini, and Grok, and a prioritised fix roadmap that updates as your work changes the citation behaviour.
For deeper context on how we measure all this, see how to measure AI visibility and inside the AI visibility audit. For the broader why, our pieces on how to get cited by ChatGPT (a companion to this post with a slightly different angle) and GEO vs SEO: what changed cover related angles.
If you would rather talk through your specific business before doing anything technical, book a free 15-minute walkthrough and we will run a live audit on your domain and explain the most impactful changes you can make today. Most businesses we audit can identify their top three fixes within ten minutes of looking at the score breakdown together.
Getting your business mentioned by ChatGPT is not a marketing trick. It is a sequence of technical and content decisions that signal to AI engines that you are a credible, structured, accessible source worth citing. The businesses that act on this now will be the default recommendations for their industries by the time their competitors catch on.