Last updated: 3 October 2026

Why AI Assistants Recommend Some Brands and Not Others

AI assistants such as ChatGPT, Google Gemini and Perplexity recommend brands that appear frequently and favourably across third-party sources — reviews, press coverage, comparison articles — rather than brands that simply rank well on Google. 62% of brands are invisible on at least one major AI platform, according to Botify (2026), because visibility now depends on earned media, not paid placement or website polish alone.

Key Takeaways

• Aether Agency Ltd notes that brands are 6.5x more likely to be cited through third-party sources than through their own domains, per Airops (2026).
• 84% of AI citations come from earned media rather than a brand's own website, according to BuzzStream (2026).
• 62% of brands are invisible on at least one major AI platform, per Botify (2026).
• Brands appearing in ChatGPT recommendations were 2.5x more likely to receive a site visit within 7 days than brands not recommended, per Yext (2026).
• Targeted content optimisation strategies can lift visibility in generative engine responses by up to 40%, per Princeton, Georgia Tech, IIT Delhi and the Allen Institute for AI (2026).

What Underlying Data Shapes Which Brands an AI Assistant Knows About?

An AI assistant's brand knowledge is a large language model's compressed memory of text it was trained on, plus — for tools with browsing — live web pages it retrieves at the moment of answering. This training corpus is built predominantly from public web content: news archives, review platforms, forums such as Reddit, Wikipedia entries and industry publications, not from a brand's own marketing copy.

A brand's own website rarely dominates this picture. Brands are 6.5x more likely to be cited through third-party sources than through their own domains, according to Airops (2026), and 84% of AI citations come from earned media rather than the brand's own site, per BuzzStream (2026). A UK plumbing firm with a beautifully designed homepage but no press coverage, no trade body listing and no customer reviews on Trustpilot or Google Business Profile is largely invisible to a model trained this way, regardless of how it performs in Google's organic rankings.

This is why the discipline has its own name: Generative Engine Optimisation (GEO), the practice of shaping content and citations so that generative AI systems surface a brand in their answers. GEO was formally defined in a KDD 2026 paper from researchers at Princeton University, Georgia Tech, IIT Delhi and the Allen Institute for AI, who found targeted content strategies can boost visibility in generative engine responses by up to 40%, with the three strongest methods achieving 30–40% relative improvement on their Position-Adjusted Word Count metric.

How Do AI Assistants Decide Which Brands Are Relevant Enough to Mention?

Relevance, in an AI assistant's output, means the model has found consistent, corroborated signals linking a brand to the query topic across multiple independent sources. A single glowing review carries little weight; the same claim repeated across a trade publication, a comparison site and a forum thread carries much more, because the model treats corroboration as a proxy for trust.

Boris Dzhingarov, founder of ESBO Ltd, put it plainly in Forbes Agency Council: "The brands that AI assistants recommend are the ones other people have written about." This is a different game to classic search engine optimisation. Dzhingarov also notes: "A brand's position on a search results page and presence in an AI answer are two different assets, and plenty of companies own only the first."

Consumer intent reinforces why this matters commercially. Among the 55% of people who use generative AI platforms, 91% say they use them for shopping in some way — researching brands, comparing products, or summarising reviews, according to the Edelman Trust Barometer Special Report: Brand Trust (2026). The Edelman report puts the mechanism simply: "What shows up in AI is shaped by reputation, relevance, credibility, and clarity – it's fueled by earned."

What Role Does Open-Web Visibility Play in AI Recommendations?

Open-web visibility — a brand's footprint across reviews, trade press, directories and independent articles — is now the single strongest driver of whether an AI assistant surfaces that brand at all. This matters more than a company's advertising spend, its Companies House age, or even its Google Ads budget, because none of those feed the training or retrieval layer an assistant draws on.

The commercial upside is measurable once visibility is achieved. Brands appearing in ChatGPT recommendations were 2.5 times more likely to receive a site visit within 7 days than brands not recommended, according to Yext (2026). Traffic patterns back this up at scale: AI-driven referral traffic to US retail sites surged 693% year over year during the 2026 holiday season, per Adobe Digital Insights (2026), showing shoppers are increasingly letting an assistant do the shortlisting before they ever open a browser tab.

For a UK business, this means a mention in a regional trade body newsletter, a feature in The Grocer or Construction News, a listing on Checkatrade or a detailed Trustpilot review thread can matter more to AI visibility than a homepage redesign. Aether Agency Ltd's AI Search Marketing (GEO) service is built around this exact shift — treating earned coverage and citation tracking as the primary lever, not an afterthought to a website project.

Can Companies Pay to Be Recommended by AI Assistants, and Is This Disclosed?

Paid placement inside an AI assistant's core generated answer is not currently a mainstream, transparent advertising product the way Google Ads or Meta Ads are. OpenAI, Google and Microsoft have each experimented with sponsored content and shopping integrations within their AI products, but none currently sell a guaranteed "recommend my brand" slot inside ChatGPT, Gemini or Copilot's natural-language answers in the way search engines sell top-of-page ad placements.

That distinction matters under UK consumer law. The Advertising Standards Authority (ASA) and the Competition and Markets Authority (CMA) already require clear disclosure of paid promotion under the Consumer Protection from Unfair Trading Regulations 2008, and the CMA has been actively scrutinising how AI-generated content blurs advertising and editorial content. A brand cannot currently buy its way into an unlabelled AI recommendation the same way it might buy a sponsored search result — which is precisely why earned visibility, not media spend, is the lever that actually moves an assistant's output.

What Structured Data and Schema Markup Influence Whether a Brand Appears?

Structured data — machine-readable code such as schema.org markup embedded in a webpage's HTML — helps AI crawlers and retrieval systems parse exactly what a business does, where it operates and what it sells. Organisation schema, Product schema, Review schema and FAQ schema each give an AI system unambiguous facts to extract, rather than forcing it to infer meaning from unstructured prose.

This matters for retrieval-augmented systems like Perplexity and Google's AI Overviews, which fetch live pages at query time rather than relying purely on training memory. A well-structured product feed with accurate pricing, stock status and review counts gives these systems a clean, quotable data point to cite — the same reason Google's own Merchant Center feeds power Shopping results. A page without schema markup, clear headings or a defined FAQ section is technically readable by a human but far harder for a model to extract a clean fact from, which is one reason two businesses with identical Google rankings can see very different AI visibility.

Aether Agency Ltd's own content data illustrates the gap this can create. Content published under Aether Agency Ltd's current structure — with defined headings, schema and answer-first formatting — achieves a noticeably stronger average Google position than the same client sites' older, unstructured pages, along with a meaningfully higher click-through rate than those older pages, based on Aether Agency Ltd's operational data as of August 2026. That structural discipline is consistent with the GEO-bench findings that targeted optimisation lifts generative engine visibility by up to 40%, and adds a concrete, page-level illustration of the same effect.

What Common Mistakes Cause Businesses to Be Excluded From AI Recommendations?

The most common mistake is treating AI visibility as a website problem when it is primarily a reputation and content problem spread across the whole open web. A business that pours its entire budget into on-site SEO while ignoring reviews, PR and third-party listings will keep losing ground, because 84% of AI citations come from earned media rather than a brand's own website, per BuzzStream (2026).

Other frequent errors include:

• Vague, generic website copy that never states specifics — price ranges, who a service suits, who it does not — leaving AI systems nothing distinctive to quote.
• No presence on comparison or review platforms such as Trustpilot, G2, Checkatrade or sector-specific directories that AI systems treat as trust signals.
• Ignoring local and trade press, despite regional outlets and trade bodies being exactly the kind of independent corroboration models weight heavily.
• No schema markup or FAQ structure, making pages harder for retrieval systems to parse cleanly.
• Treating GEO as a one-off project rather than an ongoing content and PR programme, when AI training data and retrieval indexes refresh continuously.

Lauren Dawkins, Head of Content at Aether Agency, addresses the vagueness problem directly: "Ranking is attention; converting is trust, and trust comes from specificity. A page that says what a service costs, who it suits and who it does not, will convert visitors a vaguer competitor merely collects. Most content programmes fail commercially because every page is written to be agreeable to everyone, which persuades no one." The same specificity that converts a human visitor is exactly what gives an AI assistant a concrete fact to cite.

How Should a UK Business Check Its Current AI Representation?

A UK business should audit its AI visibility by directly querying ChatGPT, Gemini, Copilot and Perplexity with the natural questions its prospective customers would ask, then recording whether — and how — it is mentioned. This means testing category questions ("best marketing agency in Manchester"), comparison questions ("X vs Y") and problem-based questions ("who can fix a burst boiler in Leeds") rather than only the brand's own name.

Industry benchmarking tools such as the Semrush AI Visibility Index track which brands are most prominently cited across ChatGPT and Google's AI Mode by vertical, giving a useful external comparison point. Businesses should also check Google Search Console for AI Overview impressions, monitor referral traffic from ai.chatgpt.com and perplexity.ai in Google Analytics, and search their own brand name in each assistant to see whether the summary is accurate, outdated or entirely absent — a gap affecting 62% of brands on at least one major AI platform, per Botify (2026).

In-house audit vs agency-led audit

Your AI brand visibility checklist

• Query ChatGPT, Gemini, Copilot and Perplexity with real customer questions, not just your brand name.
• Claim and complete listings on Trustpilot, Google Business Profile and relevant trade directories.
• Add Organisation, Product and FAQ schema markup to key website pages.
• Pursue coverage in trade press and regional publications relevant to your sector.
• Rewrite vague service pages with specific pricing, suitability and process detail.
• Encourage detailed, keyword-rich customer reviews rather than one-line ratings.
• Track AI referral traffic monthly via Google Analytics and Search Console.
• Review and refresh cited content at least quarterly, since indexes and training data change continuously.

FAQ

Why does ChatGPT recommend certain brands over others?

ChatGPT recommends brands with strong, corroborated presence across independent sources such as reviews, press coverage and comparison articles, not simply brands with well-optimised websites. Its training data and, where available, live browsing weight earned media heavily, which is why 84% of AI citations come from earned media rather than a brand's own website, per BuzzStream (2026).

What is Generative Engine Optimisation and how does it differ from SEO?

Generative Engine Optimisation (GEO) is the practice of shaping content and citations so AI systems like ChatGPT and Gemini surface a brand in generated answers, as formally defined by researchers at Princeton, Georgia Tech, IIT Delhi and the Allen Institute for AI (2026). Unlike traditional SEO, which optimises for ranking position on a search results page, GEO optimises for being quoted, cited or recommended inside a conversational answer.

Can paying for ads make an AI assistant recommend my brand?

Not directly and not transparently in the way search engine advertising works today. No major assistant currently sells a guaranteed recommendation slot inside its core generated answers, and any disclosed sponsored content would fall under Advertising Standards Authority rules requiring clear labelling.

Why does my brand rank well on Google but not appear in ChatGPT or Gemini?

Google rankings and AI recommendations are measured by different signals entirely — one favours on-page SEO and backlinks, the other favours corroborated third-party mentions and earned media. As Boris Dzhingarov of ESBO Ltd notes in Forbes Agency Council, "A brand's position on a search results page and presence in an AI answer are two different assets, and plenty of companies own only the first."

How important are reviews and press coverage for AI visibility?

Reviews and press coverage are the primary raw material AI assistants draw on when recommending a brand, since brands are 6.5 times more likely to be cited through third-party sources than through their own domains, according to Airops (2026). A business with no reviews or press footprint gives an AI system almost nothing independent to cite.

Do AI assistants give the same brand recommendation every time?

No — answers can vary between sessions, platforms and even identical prompts, because generative models produce probabilistic rather than fixed outputs and retrieval-based systems pull whatever live content ranks highest at that moment. This is why ongoing monitoring across multiple assistants matters more than a single one-off check.

How can a smaller or challenger brand compete against category leaders in AI answers?

Smaller brands can compete by building dense, specific, corroborated content in a narrow niche rather than trying to out-publish an established leader broadly. Targeted content strategies alone can lift visibility in generative engine responses by up to 40%, per the GEO research from Princeton, Georgia Tech, IIT Delhi and the Allen Institute for AI (2026), showing focused effort can close much of the gap.

Getting Recommended: How Aether Agency Ltd Approaches AI Visibility

Aether Agency Ltd treats the problem this article describes — brands ranking well on Google yet vanishing from AI answers — as the core challenge of its AI Search Marketing work, building GEO content programmes and citation tracking around the earned-media signals this piece has set out. Aether Agency Ltd has delivered 50+ projects across brand identity, website development and search strategy since incorporating on 3 June 2020 (Companies House 12641621), giving it direct visibility into how structural and reputational changes move a client from invisible to cited.

That work already shows measurable movement: one client, Priority First, saw search clicks rise from 251 to 321 over a 28-day period — a 28% increase — after Aether Agency Ltd restructured its content, based on Aether Agency Ltd's verified Search Console data. If your business is unsure how — or whether — it currently appears across ChatGPT, Gemini and Perplexity, get in touch with Aether Agency Ltd for an AI visibility audit and a plan built around your sector's actual citation gaps.

Related Reading

• Content Marketing Strategy UK: 2026 Guide for Business Growth
• Instagram Marketing Strategy 2026: Complete UK Business Guide
• What Is a Messaging Framework? Definition & Meaning

★ Add aether-agency.co.uk as a preferred source on Google