United Press · AI Visibility & Publication Choice
UK AI Visibility & GEO AgencyWhen The Answer
Never Names You.
AI visibility is simple to describe and hard to buy: ask any assistant who the best in your category is, and it names a handful of companies. That answer is assembled from published articles, so which publications carry your name decides whether you are in it. Reuters, Time, The Independent, TechCrunch, Evening Standard and around 1,400 more titles — and an honest account of which ones actually move an AI answer.
- ★Every title namedYou see the publication list before anything is written
- ◆Indexed, not just listedA title an assistant cannot read is worth nothing
- ●No invented metricsNobody can measure “AI visibility” as a score
- ↻Live link for every articleA permanent, indexed URL an assistant can actually read
On This Page
- What “AI visibility” actually means
- The surfaces: ChatGPT, AI Overviews, AI Mode, Perplexity
- The publications worth naming
- How an assistant decides who to name
- What actually correlates with being named
- How to pick publications for AI visibility
- Why competitors appear and you do not
- What this tier cannot promise
- GEO, AEO and AI SEO explained
- Being described correctly
- AI visibility FAQ
Read this first
What “AI Visibility” Actually Means, and What It Does Not
This is the most oversold phrase in the industry right now, so here is the plain version before the list.
There is no AI visibility score
Nobody can measure how visible you are to an assistant, because there is no index to query and no dashboard that reports it. Any agency showing you an AI visibility score built it themselves.
No publication is certified for AI visibility
There is no programme, no badge and no approval status. When a media list ticks every row “AI visible” — and they do — that tick is a sales label, not a measurement. We will not repeat it back to you as fact.
What genuinely helps AI visibility
Being written about on sites that are indexed, established, and already cited by other sites. Assistants build answers from what they can read and what other sources corroborate. That is the whole mechanism.
Being named beside your category
An assistant answering “best X” needs your name to sit inside an article about X. A profile piece that never states your category rarely surfaces for the question you care about.
Indexed is the minimum bar for AI visibility
A title an assistant cannot crawl is worth nothing here, whatever its follower count or claimed reach. Indexing is the one thing on this page that can actually be checked, and we check it.
What we tell you honestly about AI visibility
Which titles are strong, which are filler, and that no one can promise you a citation in an AI answer. What can be promised is where the article goes and that it is indexed when it lands.
Where the answers appear
AI Visibility: The Surfaces You Are Trying to Appear In
People say “AI search” as though it is one thing. It is at least six, they behave differently, and a name that shows up in one may be missing from the next.
ChatGPT
The one most buyers use by name. It answers partly from training and partly by searching live, and it correlates less strongly with conventional authority than the others — which makes it the most reachable surface for a smaller brand.
Google AI Overviews
The generated summary above the blue links. It draws heavily on pages already ranking for the query, so conventional UK search visibility and AI Overview presence tend to move together.
Google AI Mode
Google’s full conversational search. It tracks established brand signals more closely than any other surface, behaving like a consensus engine — hardest to break into, most durable once you are there.
Perplexity
Answers with visible citations, which makes it the easiest place to audit your own presence. If you want to see whether coverage is being read, start here.
Claude and Copilot
Both answer from a mix of training and live retrieval. Coverage on established, indexed publications is what reaches them; nothing platform-specific is required.
Assistants inside other products
Shopping assistants, in-app search and voice assistants increasingly summarise the same open web. The work that gets you into one generally serves the rest.
The list, in the open
The Publications Worth Naming for AI Visibility
The full catalogue runs to roughly 1,400 titles, and most of it is regional and niche filler that we would not recommend. These are the ones with the standing to affect an answer. Every title on your shortlist is named, with its indexing status, before you approve anything.
International Newswires & Majors
The titles other publications quote, which is what makes an assistant treat them as corroboration.
UK Nationals
UK mastheads, which matter disproportionately when the question is phrased around Britain.
UK Regional Press
Large regional mastheads. Individually modest, collectively the densest UK coverage on the list.
Technology
Where an assistant looks when the question is about software, hardware or a technical category.
Business & Finance
Titles that carry weight in investor, procurement and B2B questions.
US Metro & Regional
City mastheads that dominate geo-qualified questions such as “best in Los Angeles”.
Entertainment, Music & Lifestyle
Culture titles, which assistants lean on heavily for people, brands and creative categories.
International Editions
Regional editions covering the Middle East, Africa, Asia Pacific and Europe.
How it works
How an Assistant Decides Who to Name in an AI Visibility Answer
Every assistant does this slightly differently, but the shape is the same and it is not mysterious.
AI visibility starts with what it can crawl
Indexed web pages, overwhelmingly. Anything behind a login, a paywall or a platform that blocks crawlers is largely invisible to it, whatever its audience size.
It looks for the category, not the brand
The question is “best X”, so it searches for articles about X. Your name has to appear inside content about your category, not only in coverage about you.
It weighs corroboration
A name appearing across several independent, established sources is treated as more reliable than the same name on one site. This is why one big placement rarely changes an answer on its own.
It prefers sources that look editorial
Assistants are increasingly trained to discount self-published pages, press releases and thin sites. A company describing itself carries less weight than a publication describing it.
It favours recent and dated material
Questions implying currency pull recent sources. A dated 2026 article outperforms an undated one, which is why round-ups get refreshed annually.
Social posts add almost no AI visibility
Most social platforms block crawlers, so posts are largely unreadable to assistants. Reach on social is real; AI visibility from it is mostly not.
What the evidence says
What Actually Correlates With AI Visibility
Ahrefs studied around 75,000 brands across ChatGPT, Google AI Mode and AI Overviews and measured which signals track with being mentioned. The findings are unusually clear, and they contradict most of what is being sold.
YouTube mentions — the strongest signal (~0.74)
Brand names appearing in video titles, descriptions and transcripts correlated more strongly than anything else measured. Video is the most under-used route into AI answers by a wide margin.
Branded web mentions — close behind (0.66–0.71)
Your name appearing in published content across the web. This is what media placement produces, and it is the second strongest correlation in the study.
Branded anchors (0.51–0.63)
Links where the anchor text is your brand name. Related to mentions, and weaker than plain mentions on their own.
Branded search volume (0.35–0.47)
How many people search your name. Moderate, and largely a downstream effect of the two signals above.
Backlinks and URL rating — very weak
The study found very weak correlations for both. This is the single most important finding for anyone being sold links as an AI-visibility product.
More content — almost no relationship (~0.19)
Site page count barely registered. Publishing more pages on your own site is close to the least effective thing you can do for this.
Domain rating — modest (0.27–0.33)
A high-authority domain helps a little. Far less than the industry’s obsession with domain metrics would suggest.
What AI visibility means in practice
Get mentioned in published content, on video and on the open web, by name. That is media relations. It is not link building, and it is not content volume.
And why the platforms differ
AI Mode tracked brand authority most closely; ChatGPT least. A smaller brand realistically targets ChatGPT and Perplexity first, and reaches AI Mode later or not at all.
Choosing titles
How to Pick Publications for AI Visibility
Six rules that decide whether a placement does anything. They are the same rules we apply to your shortlist.
Check it is indexed, or there is no AI visibility
Search a distinctive phrase from one of its recent articles. If Google has never seen the site, no assistant will read it either. This is the single fastest way to disqualify a title.
Check whether anyone quotes it
Do other publications cite its reporting? Corroboration is what assistants weigh, and a site nobody else references adds a name without adding confirmation.
Match the title to the category
A general news site publishing a piece about you helps less than a sector title publishing the same piece, because the second one sits inside the topic an assistant is searching.
Match the geography to the question
If your buyers ask about the UK, UK mastheads and regionals carry more weight in that answer than larger US titles that never mention Britain.
Prefer titles that keep archives live
Some sites rotate content out after months. A placement that disappears takes its citation with it, so we favour publications with stable, permanent archives.
Ignore the AI visibility tick
Every row on every media list in this industry is ticked AI-visible, including sites with a domain authority of three. Judge the publication, not the label somebody typed next to it.
The diagnosis
AI Visibility: Why Competitors Appear and You Do Not
It is almost never that the assistant dislikes you. It is one of six things, and most are fixable.
Nobody has written about you in your category
The commonest cause by a distance. An assistant answering “best X in the UK” searches for articles about X. If your name only appears in coverage about your funding round, it is not in the pool it is drawing from.
You are mentioned, but only on your own site
Self-published pages are heavily discounted. A company describing itself is not corroboration, and corroboration is the mechanism.
Your mentions are on sites nothing else quotes
One placement on a thin site adds a name without adding confirmation. Assistants weigh how many independent sources agree, not how many URLs exist.
Your competitors are in the round-ups and you are not
Best-of articles already contain the comparison an assistant is looking for. Being absent from those lists is the fastest way to be absent from the answer — our listicle placement page covers that route.
Nothing recent, so no AI visibility
Questions implying currency pull recent sources. Coverage from three years ago carries much less weight than a dated piece from this year.
You are described in the wrong words
If published articles call you a “marketing consultancy” and buyers ask for a “PR agency”, you will not surface. The vocabulary in the coverage has to match the vocabulary in the question.
Being honest about it
What AI Visibility Work Cannot Promise
Four things we will not tell you, and the reason each one gets sold anyway.
A guaranteed AI visibility citation
No one controls what a model outputs, and the same question can produce different answers on different days. An article is a real, permanent asset; a citation is a probability, and anyone selling one as a certainty is guessing.
An AI visibility score or percentage
There is no data source for it. Scores in AI-visibility dashboards are constructed from a handful of sample prompts, which means they measure the sample rather than the model.
A fixed timescale
Models are retrained and their search layers refresh on schedules nobody outside those companies knows. Content published today may be picked up in weeks or not at all.
That volume alone works
A hundred placements on thin sites move less than a handful on titles other publications actually quote. Buying the long tail of any media list is the most common way money gets wasted here.
What we do promise
Named publications agreed before anything is written, articles that state your category clearly, indexing checked on delivery, and a live link for each one.
How to judge AI visibility yourself
Ask the assistants your buyers use, in your buyers’ words, before we start and again in a few months. It is a rough measure, but it is an honest one and it costs nothing.
The vocabulary
AI Visibility Vocabulary: GEO, AEO and AI SEO Explained
Three acronyms for roughly the same idea, sold at very different prices.
GEO — generative engine optimisation
Getting named in answers produced by generative assistants. In practice it means earning mentions on sources those assistants read and trust.
AEO — answer engine optimisation
The same idea framed around answer boxes and featured snippets as well as chatbots. The tactics overlap almost completely with GEO.
AI SEO
A looser term covering both, plus using AI tools to do conventional SEO. Worth asking anyone selling it which of those two things they actually mean.
Citation vs mention
A citation links back to the source; a mention just names you. Assistants use both, but a linked citation is the one you can point to and verify.
Share of voice
How often your name appears in answers across a set of questions, compared with competitors. Useful as a rough direction of travel, not as a metric anyone can measure precisely.
Retrieval vs training data
Assistants either recall what they learned during training or fetch live pages when answering. New coverage reaches the second route quickly and the first one slowly, if at all.
Beyond being mentioned
AI Visibility Is Also Being Described Correctly
Appearing in the answer is half of it. The sentence the assistant uses about you is the other half, and it comes straight from published sources.
AI visibility copy is lifted, not invented
Assistants characterise you using the words published articles use. If the coverage is vague, the description is vague. This is why what an article says about you matters as much as where it runs.
Category language decides your questions
Being called a “UK PR agency” surfaces you for one set of questions; “communications consultancy” for another. Pick the words your buyers actually use and get them into the coverage.
Wrong facts propagate
An out-of-date figure or a wrong location repeated across a few articles becomes the version assistants report. Corrections are slow, so it is worth getting right the first time.
Founders and companies are separate entities
An assistant can know your founder well and your company barely at all. If the company name is the one buyers search, it needs to appear in its own right.
Check your AI visibility now
Ask three assistants to describe your company, in your buyers’ words, before commissioning anything. What comes back tells you whether the problem is absence or description.
Consistency across sources builds AI visibility
The same category, the same claims and the same name spelling across several publications is what turns a scattered set of mentions into one confident answer.
Questions
AI Visibility, GEO & AEO: Your Questions Answered
Straight answers on ChatGPT, AI Overviews, AI Mode and Perplexity — what the evidence supports, and what nobody can promise.
What does AI visibility mean?
Being named when someone asks an assistant a question in your category. Assistants build those answers from published articles, so which publications carry your name decides whether you appear in one.
What is an AEO agency?
An answer engine optimisation agency works on whether your brand is selected, cited and described in AI-generated answers rather than on where you rank in a list of blue links. In practice most of the work is earning mentions on sources those systems read, which is media relations aimed at a new destination.
Is GEO the same as AEO?
Near enough in practice. GEO frames it around generative assistants, AEO around answer boxes and snippets as well. The tactics overlap almost entirely, and the price difference between agencies selling each is not explained by the method.
Why do my competitors appear in AI answers and I do not?
Usually one of six things: nothing published about you in your category, mentions only on your own site, mentions on sites nothing else quotes, absence from the round-ups your competitors are in, nothing recent, or coverage describing you in words your buyers do not use.
Can this improve my ChatGPT visibility?
It is the surface most likely to respond. Research across roughly 75,000 brands found ChatGPT correlates least strongly with conventional authority signals, which makes it the most reachable of the answer surfaces for a smaller or newer brand. Nobody can guarantee a mention on any of them.
Can this improve my Google AI Overviews visibility?
AI Overviews lean heavily on pages already ranking for the query, so UK search visibility and AI Overview presence tend to move together. Coverage on established, indexed publications helps both.
What is Google AI Mode and is it different?
It is Google’s full conversational search. In the same research it tracked established brand signals more closely than any other surface, behaving like a consensus engine. It is the hardest to break into and the most durable once you are in it.
Can you guarantee ChatGPT will mention my brand?
No, and nobody honestly can. No one controls what a model outputs and the same question can return different answers on different days. What can be promised is where the article is published and that it is indexed when it lands.
Is there such a thing as an AI visibility score?
Not a real one. There is no index to query and no dashboard that reports it, so any score you are shown was constructed by the agency showing it, usually from a handful of sample prompts. It measures the sample, not the model.
Are these publications certified as AI-visible?
No. There is no certification programme, no badge and no approval status anywhere. Every row on every media list in this industry is ticked AI-visible, including sites with a domain authority of three. Judge the publication instead.
How do AI assistants choose which sources to cite?
They read indexed pages, search for the category rather than your brand, weigh how many independent sources corroborate a name, prefer editorial sources over self-published ones, and favour recent, dated material.
Which publications actually influence AI answers?
The ones other publications quote. Reuters, Time, USA Today, Business Insider, The Independent, The Times, TechCrunch and similar carry weight because they are widely cited elsewhere, which is what corroboration means in practice.
How many publications are on your list?
Roughly 1,400 in the full catalogue, and we would not recommend most of them. A large part is regional and niche filler. The titles worth naming are grouped by sector on this page, and your shortlist is always named before you approve.
How do you measure AI visibility?
By asking the assistants your buyers use, in your buyers’ words, before we start and again later, and by recording what comes back. It is rough, and we say so. We do not produce a visibility score, because there is no data source for one.
How long does GEO or AEO take?
Nobody outside the model companies knows their retraining and refresh schedules. Live-retrieval surfaces such as Perplexity can reflect new coverage within weeks; training-based recall may never pick it up. Any agency quoting you a fixed timescale is inventing it.
Do you guarantee AI mentions or citations?
No. No one controls what a model outputs and the same question returns different answers on different days. The article is the deliverable and it is permanent; the citation is a probability.
What correlates most strongly with being mentioned by AI?
YouTube mentions, at roughly 0.74 — your brand name in video titles, descriptions and transcripts. Branded web mentions followed at 0.66 to 0.71. Both are mention-based signals, which is why placement works and link buying does not.
Do backlinks help AI visibility?
Barely. The 75,000-brand study found very weak correlations for backlinks and URL rating. This matters because links are still the most commonly sold AI-visibility product, and the evidence does not support it.
Does domain authority matter for AI visibility?
Modestly. Domain rating correlated at roughly 0.27 to 0.33 — real but far weaker than the industry’s focus on domain metrics implies. A relevant mention on a mid-authority title beats a vague one on a big domain.
Does publishing more content on my site help?
Almost not at all. Site page count correlated at around 0.19, close to nothing. Publishing more pages about yourself is one of the least effective things you can do for AI visibility, however often it is recommended.
Does YouTube really help AI visibility?
It was the single strongest correlation in the study, and it is the most under-used route by a wide margin. We do not sell video, but if you are choosing where to spend, it deserves more attention than it gets.
Is one big placement enough?
Rarely. Assistants weigh corroboration, so a name appearing across several independent, established sources holds up better than the same name on one site, however impressive that one site is.
Does being in a listicle help more than a standalone article?
Usually yes, for this purpose. A round-up already contains the comparison an assistant is looking for, which is why best-of articles get quoted so heavily. Our listicle placement page covers that route specifically.
Do social media posts help AI visibility?
Very little. Most social platforms block crawlers, so posts are largely unreadable to assistants. Social reach is real and worth buying for its own reasons, but not for this one.
Do press releases get cited by AI?
They can be read, but they are discounted. A press release is a company describing itself, and assistants increasingly weight independent sources more heavily. Distribution supports a record; it rarely creates a recommendation.
How long does it take to see a change?
Nobody outside the model companies knows their retraining and refresh schedules. Content published today may be picked up in weeks or not at all. Any agency quoting you a timescale is inventing it.
How do I measure whether it worked?
Ask the assistants your buyers use, in your buyers’ words, before starting and again a few months later. It is rough, but it is honest and it costs nothing. Treat any polished dashboard with more suspicion, not less.
What is GEO?
Generative engine optimisation: getting named in answers produced by generative assistants. In practice it means earning mentions on the sources those assistants read and trust, which is conventional media relations aimed at a new destination.
What is the difference between GEO, AEO and AI SEO?
Very little in practice. GEO frames it around chatbots, AEO around answer boxes and snippets, and AI SEO covers both plus using AI tools for ordinary SEO. Ask anyone selling AI SEO which of those they actually mean.
Does the publication need to link to me?
A linked citation is better, because it is verifiable and it helps conventional search too. But assistants read names and context as well as links, so an unlinked mention in a strong article still has value.
Does it matter if the article is dated?
Yes. Questions that imply currency pull recent sources, so a dated current-year article outperforms an undated one and considerably outperforms an old one.
Will this help my Google rankings too?
Often yes, as a side effect, because the same qualities matter: indexed pages on established sites. But if UK search rankings are the main goal, say so and we will weight the shortlist differently.
How do assistants describe my company, and can that be changed?
They lift the characterisation from published sources. Change what the coverage says and the description follows, though slowly. Ask three assistants to describe you before commissioning anything — the answer tells you whether your problem is absence or description.
Will an AI assistant repeat wrong information about me?
It can. An out-of-date figure or wrong location repeated across a few articles becomes the version that gets reported. Corrections are slow, which is an argument for getting the first wave of coverage accurate.
Which AI assistant is easiest to appear in?
On the evidence, ChatGPT and Perplexity. Perplexity shows its citations, which makes it the best place to audit whether your coverage is being read at all. AI Mode is the hardest.
Is AI visibility work different from SEO?
It overlaps but is not the same. SEO optimises pages you own to rank; this optimises what other people publish about you so you get named. The signals that matter are mentions rather than links, which is why the two disciplines are drifting apart.
Do I need to be in the UK results specifically?
If your buyers ask questions phrased around Britain, yes. UK nationals and regional mastheads carry more weight in those answers than larger US titles that never mention the UK, and we weight the shortlist accordingly.
What is the difference between a citation and a mention?
A citation links back to the source and is verifiable. A mention just names you. Assistants use both, but a linked citation on a surface like Perplexity is the one you can actually point at.
Is buying placements safe after the Google spam updates?
Buying links to manipulate rankings is a named spam-policy violation and the risk is real. We place written articles on named publications and report them as placements. We also tell you which titles are weak rather than selling you the long tail.
Why do you tell me which titles are filler?
Because you will find out anyway, and because a hundred placements on thin sites move less than a handful on titles other publications quote. Selling you volume would be easier and would work less well.
Can you do this for a UK-focused business?
Yes, and the shortlist changes. If your buyers ask questions phrased around Britain, UK nationals and regional mastheads carry more weight in that answer than larger US titles that never mention the UK.
What sectors does this work best for?
Anything with a comparison question behind it — software, professional services, finance, health, education, consumer products. If people ask an assistant which one to choose, this applies.
What do you need from me to start?
Your category in your buyers’ words, the questions you want to appear for, and the markets that matter. We come back with a named publication shortlist and an honest view of what each one is worth.
How is this different from your other AI pages?
Those explain the method — how AI search works, what GEO is, how to rank in AI answers. This page is the inventory: which publications we can actually place you in, and how to judge them.
Go deeper
AI Search, GEO & Placement
This page is the inventory. These cover the method, and the other routes into an AI answer.
How AI search works
The mechanism, explained from the beginning
GEO agency
Generative engine optimisation as a service
AI SEO agency
Where AI visibility and conventional SEO overlap
AI PR agency
Media relations aimed at AI answers
PR for AI startups
Coverage for companies building AI, not seeking it
Listicle & best-of placement
The article format assistants quote most
The full media list
Every placement tier, by sector and region
News distribution
Announcements on named news sites within days
Let’s Talk
Get the Named Publication List
Tell us what has happened, when it needs to be live and which sectors matter. We will write the announcement and come back with the named list of sites for you to approve. No obligation.