Seven signals determine whether AI search engines recommend your business: positioning clarity, FAQ structure, schema markup, an AI guide file, review freshness, third-party authority, and entity consistency. Building these signals is different work from traditional SEO, and most businesses are missing two or three of them.
How to get recommended by AI search engines
Seven specific signals determine whether AI search engines recommend your business. Build all seven and ChatGPT, Perplexity, and Google's AI Overviews have what they need to name you confidently. Miss two or three and they skip you, even if you rank well on Google.
Most business owners assume AI recommendations follow Google rankings. They do not. AI engines make independent decisions, drawing from sources that traditional search optimisation does not touch: third-party review platforms, directory listings, structured data, and guide files your website either has or does not.
The good news is that most of these signals are straightforward to build. The bad news is that most businesses have not built them. According to Conductor's analysis of 21.9 million queries, AI Overviews now appear in 25% of Google searches, up from 13% twelve months ago. Ahrefs found that AI features reduced click-through rates for top-ranking content by 58%. The opportunity is real and it is moving fast.
This post covers all seven signals, what the research shows about each one, and a concrete sequence for building them. If you want the diagnostic version first, the post on why ChatGPT doesn't recommend your business walks through the seven gaps and which ones to fix first.
Key Takeaways
- AI engines make recommendation decisions on seven signals, most of which are independent of your Google search ranking.
- Positioning clarity is the first signal: AI engines skip businesses whose homepage headline does not clearly state who they are and what they do.
- Third-party review platform presence produces a 3x increase in ChatGPT citation likelihood — review breadth matters as much as recency.
- Content updated within 30 days receives 3.2x more AI citations than static content, making freshness a recurring advantage, not a one-time task.
- An llms.txt guide file costs 30 minutes to write and tells AI tools which pages on your site actually matter — the vast majority of small business sites do not have one.
Why does strong SEO performance fail to produce AI search recommendations?
Google rankings and AI recommendations are built on different foundations. Google evaluates backlinks, keyword relevance, and domain authority. AI engines evaluate whether they can identify your business, verify it through outside sources, and extract a clean, confident answer from your content.
The clearest example is local recommendations. Research into ChatGPT's local recommendation behaviour shows that Foursquare provides over 70% of the data signal, not Google Maps. A business that has optimised for Google Business Profile but neglected Foursquare and Yelp may rank prominently in Google's local pack while being invisible to ChatGPT.
The second failure mode is content structure. AI engines do not rank pages — they extract answers from them. A well-written page with no FAQ structure, no schema markup, and no clear answer sections gives an AI engine nothing reliable to lift. The engine moves on and finds the information somewhere it can extract cleanly.
Princeton University's GEO research, published at ACM KDD 2024, found that content optimised with traditional SEO tactics, including high keyword density, can actively reduce AI visibility. AI engines interpret keyword repetition differently from how ranking algorithms do, and they weight information density, citation quality, and structural extractability much more heavily.
What does the research say about what actually drives AI recommendations?
SE Ranking's study of 2.3 million pages found that domain traffic is the strongest single predictor of AI Mode citations. High-traffic sites earn approximately three times more AI citations than low-traffic sites, because traffic is a proxy for the compounding trust signals, including content quality, brand recognition, and review presence, that AI engines are actually evaluating.
Third-party platform presence is the most immediately actionable finding from available research. Businesses with profiles on Trustpilot, G2, Capterra, or Yelp have three times higher odds of being chosen by ChatGPT as a source, compared to businesses without such presence. The platforms function as trust anchors. Multiple independent sources corroborating a business's existence and quality increase AI recommendation confidence.
Content freshness carries a similar magnitude of effect. Research on AI citation patterns found that pages updated within 30 days receive 3.2 times more AI citations than static pages. Among ChatGPT's most-cited pages, 76.4% were updated within the previous month. This makes a monthly content refresh programme one of the highest-leverage recurring investments a business can make.
Princeton's GEO study found that adding statistics improved AI visibility by approximately 41%, expert quotations by roughly 28%, and citing authoritative sources by around 30%. Structured heading hierarchies appear in 68.7% of cited pages. The pattern is consistent: AI engines reward content that is specific, structured, and externally grounded.
The Seven-Signal AI Recommendation Framework
AI recommendation decisions follow a consistent pattern. Before an AI engine recommends a business, it is implicitly running through seven checks. The Seven-Signal AI Recommendation Framework maps those checks to signals you can build.
Signal 1: Positioning clarity. Can the AI engine identify in one sentence who you are and what you do?
The headline at the top of your homepage is the first thing AI tools read. A vague headline like "Your trusted local partner" gives an AI engine nothing to work with. A specific headline like "Employment lawyers in Melbourne specialising in unfair dismissal claims for small business owners" gives it everything. This is the fastest fix available and the most commonly missed.
Signal 2: FAQ structure. Does your site have customer-language questions answered in extractable format?
AI engines produce answers in question-and-answer format. When your website has content in the same format, the engine can pull it directly into its answer. The questions must be written the way a real customer types them, not the way a business describes its own services. Five to ten well-written FAQ entries, published on a page AI tools can find, address this signal directly.
Signal 3: Schema markup. Does your website have the background code that tells AI engines who you are in machine language?
Schema markup is structured data that describes your business to automated systems without requiring them to interpret your visible content. The four most useful types are Organisation schema (name, industry, contact details), LocalBusiness schema (address, service area), Service schema (what you offer), and FAQPage schema (which marks your FAQ section as machine-readable). Princeton's GEO research found that structured data appears in 61% of cited pages.
Signal 4: AI guide file. Does your site have an llms.txt file that tells AI tools which pages matter?
An llms.txt file is a short text document, placed on your website, that tells AI crawlers which pages are most important and how to interpret your content. AI tools including Perplexity check for it when visiting your site. Without it, they make their own decisions about which pages to prioritise. Writing one takes approximately 30 minutes and the format guide is available at llmstxt.org. It is one of the fastest available improvements and the majority of small business websites do not have one.
Signal 5: Review freshness. Do you have recent, specific reviews on Google and at least one third-party platform?
AI engines use reviews as an outside verification signal. Research found that businesses with profiles on Trustpilot, Yelp, and similar platforms have three times higher odds of being cited by ChatGPT. Complete Google Business Profile data correlates with 2.8 times higher frequency of AI search appearances. Reviews posted in the last 12 months, with specific detail about the service and outcome, carry far more weight than older or generic entries.
Signal 6: Third-party authority. Do credible external websites mention your business?
If you search your business name and it only appears on your own website, AI engines have only your own word to go on. That is not enough for a confident recommendation. Three to five mentions on credible external sites, including an industry directory listing, a professional association profile, or a supplier reference, carry more weight than dozens of generic link-building entries. The standard is credibility, not volume.
Signal 7: Entity consistency. Is your business information identical across every platform where you appear?
Name, address, phone number, website URL, and service description must match precisely across your website, Google Business Profile, Foursquare, Yelp, and any directory where you are listed. Inconsistency reduces AI confidence in the recommendation. When the same business is described differently across sources, AI engines treat it as lower-confidence and often skip it in favour of a competitor with clearer entity data.
Each signal is independent but compounds with the others. Businesses that appear consistently in AI recommendations typically have all seven. Most businesses missing from AI recommendations have two or three gaps, not all seven.
How do you build all seven signals in the next 30 days?
Step 1: Fix your homepage headline (Day 1, free)
Rewrite the main headline on your homepage so it names who you are, what you do, who you serve, and where you work, in plain language. This costs nothing and addresses the single most common reason AI engines skip businesses.
Step 2: Write and publish a FAQ section (Days 2-3, free)
Write five to ten questions your customers genuinely ask before buying from you. Answer each one directly and completely. Publish them on your website. Write the questions the way a customer types them, not the way you would describe your service.
Step 3: Add schema markup (Days 3-7, low cost)
Ask your web developer to add Organisation schema, LocalBusiness schema if location is relevant, and FAQPage schema to your FAQ section. If your site runs on WordPress with Yoast SEO or RankMath, much of this is already built in and needs configuration, not custom code. Run a free AI visibility scan to see which schema types are currently missing from your site.
Step 4: Write your llms.txt file (Day 3, 30 minutes, free)
List your key pages with a one-line description of each. Follow the format at llmstxt.org. Upload the file to your website's root directory. This is done in one afternoon and the majority of your competitors have not done it.
Step 5: Request recent reviews (Week 2, free)
Ask three to five recent customers to leave an honest, specific review on your Google Business Profile and, if relevant, on Yelp or an industry-specific platform. Specific detail matters more than star rating. A review describing exactly what you helped with gives AI engines far more to work with than a five-star entry with no text.
Step 6: Claim directory listings and third-party profiles (Week 2-3)
Identify the one or two directories most relevant to your industry and create or claim a complete profile on each. Then confirm your profile on Foursquare, Yelp, and Bing Places. Complete profiles with accurate, current information outperform skeletal ones every time.
Step 7: Audit and fix entity consistency (Week 3-4)
Search your business name across every platform where you appear. Look for mismatches in name, address, phone number, and website URL. Correct every inconsistency. Entity ambiguity, meaning inconsistent business information across platforms, is the most consistent gap identified in AI visibility analysis. Fixing it is often one afternoon of work with a high and lasting return.
Frequently asked questions about getting recommended by AI search engines
How long does it take to get recommended by AI search engines after making these changes?
Fast fixes like schema markup, an llms.txt file, and a FAQ section start showing up within days to a few weeks. Trust signals like reviews and third-party mentions take weeks to months to compound. Most businesses see measurable improvement in AI recommendations within 30 to 60 days of addressing the top two or three signals.
Do I need to fix all seven signals to appear in AI recommendations?
No. The two most common gaps are positioning clarity (the AI cannot tell who you are) and third-party authority (no outside sources confirm you exist). Fix those first. The remaining five compound the effect over time. Most businesses missing from AI recommendations have two or three gaps, not all seven.
Does my Google ranking affect whether AI engines recommend me?
Partly, but not in the way most people expect. Google rankings and AI recommendations share some inputs, including content quality and structured data. But AI engines also draw heavily from sources Google does not rank, including Foursquare, Yelp, industry directories, and third-party review platforms. A page-one Google ranking does not guarantee AI recommendations.
What is the difference between this and traditional SEO?
Traditional SEO helps you rank on a results page where a user chooses which link to click. AI recommendation work helps you get named in a single synthesised answer where the engine picks for the user. The signals overlap on content quality and site structure but diverge significantly on trust format, content extractability, and third-party platform presence.
What is an llms.txt file and does my business actually need one?
An llms.txt file is a short text document you place on your website that tells AI crawlers which pages are most important and how to interpret your content. AI tools including Perplexity check for it when visiting your site. Without it, they make their own decisions about which pages to prioritise — and those decisions are often wrong for smaller sites. It takes about 30 minutes to write and is one of the fastest improvements available. The common questions about AI visibility page covers the format in more detail.
The bottom line
Most businesses are invisible to AI search engines for the same two or three reasons. The headline is too vague. There are no outside sources confirming the business exists. The content has no structure AI tools can extract from.
Seven signals fix that. Most of them cost nothing but time. The businesses that appear consistently in AI recommendations made these structural changes early, before their competitors thought to.
Run a free AI visibility scan to see exactly which of the seven signals your site currently has and which ones are missing. Five minutes, no credit card, clear action plan.
<script type="application/ld+json"> { "@context": "https://schema.org", "@type": "Article", "headline": "How to get recommended by AI search engines", "description": "Seven specific signals determine whether AI search engines recommend your business. Here is what each one is and how to build it.", "author": { "@type": "Person", "name": "Linda Cunningham", "url": "https://getrecommended.io/founder" }, "publisher": { "@type": "Organization", "name": "GetRecommended.io", "logo": { "@type": "ImageObject", "url": "https://getrecommended.io/icon.svg" } }, "datePublished": "2026-04-27", "dateModified": "2026-04-27", "mainEntityOfPage": { "@type": "WebPage", "@id": "https://getrecommended.io/blog/how-to-get-recommended-by-ai-search-engines" } } </script> <script type="application/ld+json"> { "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [ { "@type": "Question", "name": "How long does it take to get recommended by AI search engines after making these changes?", "acceptedAnswer": { "@type": "Answer", "text": "Fast fixes like schema markup, an llms.txt file, and a FAQ section start showing up within days to a few weeks. Trust signals like reviews and third-party mentions take weeks to months to compound. Most businesses see measurable improvement in AI recommendations within 30 to 60 days of addressing the top two or three signals." } }, { "@type": "Question", "name": "Do I need to fix all seven signals to appear in AI recommendations?", "acceptedAnswer": { "@type": "Answer", "text": "No. The two most common gaps are positioning clarity (the AI cannot tell who you are) and third-party authority (no outside sources confirm you exist). Fix those first. The remaining five compound the effect over time. Most businesses missing from AI recommendations have two or three gaps, not all seven." } }, { "@type": "Question", "name": "Does my Google ranking affect whether AI engines recommend me?", "acceptedAnswer": { "@type": "Answer", "text": "Partly, but not in the way most people expect. Google rankings and AI recommendations share some inputs, including content quality and structured data. But AI engines also draw heavily from sources Google does not rank, including Foursquare, Yelp, industry directories, and third-party review platforms. A page-one Google ranking does not guarantee AI recommendations." } }, { "@type": "Question", "name": "What is the difference between this and traditional SEO?", "acceptedAnswer": { "@type": "Answer", "text": "Traditional SEO helps you rank on a results page where a user chooses which link to click. AI recommendation work helps you get named in a single synthesised answer where the engine picks for the user. The signals overlap on content quality and site structure but diverge significantly on trust format, content extractability, and third-party platform presence." } }, { "@type": "Question", "name": "What is an llms.txt file and does my business actually need one?", "acceptedAnswer": { "@type": "Answer", "text": "An llms.txt file is a short text document you place on your website that tells AI crawlers which pages are most important and how to interpret your content. AI tools including Perplexity check for it when visiting your site. Without it, they make their own decisions about which pages to prioritise. It takes about 30 minutes to write and is one of the fastest improvements available." } } ] } </script>
