Who this is for:This segment splits: established media brands and digital publishers, where citations are existential and budgets exist, are a strong fit for Prefer's done-for-you Managed service. Solo creators and small blogs are best served running these plays in-house with the free checker below.
Why AI search decides publishers and creators outcomes
8% vs 15%
click rate with vs without an AI summary
Pew analyzed 68,879 Google searches from 900 US adults: with an AI summary, people clicked a result in 8% of visits, versus 15% without. The summary keeps the reader.
1%
of AI-summary views produce a source click
In the same Pew data, users clicked a link inside the AI summary itself in just 1% of visits. Being cited in the summary rarely sends a click, so the citation is the reach.
30.3%
Reddit's share of ChatGPT source citations
In our Reddit study across 25 queries, Reddit was 30.3% of all source citations ChatGPT logged, more than double Wikipedia (13.8%). A UGC publisher is the single biggest cited source.
100M+
AI citations analyzed by Semrush
Semrush tracked over 100 million citations across ChatGPT, Google AI Mode and Perplexity for thirteen weeks; Reddit, Wikipedia, LinkedIn, YouTube and Forbes led, and the order swung sharply by engine.
The buyer prompts that decide publishers and creators
| Prompt family | An example buyer asks | What wins the citation |
|---|---|---|
| Recommendation and reviewReader wants a buying pick | best [product] 2026 | Be the structured, current authority review AI lifts |
| ExplainerReader wants to understand | what is [topic] | A clear, quotable explainer that competes with Wikipedia in your niche |
| News and current eventsReader wants the latest | what happened with [event] | Timely, primary reporting AI and Reddit both surface |
| ComparisonReader is choosing | [X] vs [Y] | Your own honest comparison, the format AI loves to quote |
| Creator credibilityReader is vetting the source | who is [creator] | Consistent bylines and an entity footprint on Wikipedia and LinkedIn |
| RoundupReader wants a curated list | top [things] in [category] | Maintained, dated listicles with clear criteria |
Example prompts are illustrative of each family; run your own category, rivals and personas to build the real set.
Which engines matter for publishers and creators, and why
- Primary
ChatGPTLargest reach and a Reddit, Wikipedia and Forbes publisher mix. Our vertical pulls were run on ChatGPT and returned the category-authority reviewers you compete with. - Primary
Google AI OverviewsThe Pew click-loss finding is a Google-AI-Overviews result and the direct threat to publisher traffic. YouTube and the reviewers lead its citations. - Secondary
PerplexityA citation-forward engine that names publishers prominently (Reddit, LinkedIn, NIH), so each citation is worth more to a media brand. - Secondary
GeminiRides Google's index of the same publisher set, so the structure and footprint work for Google AI Overviews carries over.
Who AI reads for publishers and creators answers
reddit.comCommunity / UGCThe single most-cited source in our study and Semrush's. Your content being discussed on Reddit is often how AI finds it, and the peer you compete with for attention.
en.wikipedia.orgReferenceTop on ChatGPT, near-absent on Google AI Mode. The definitional authority you compete with in your niche.
youtube.comVideoTop on Google and the core creator surface. Video is a citation channel, not just distribution.- Category-authority reviewersEditorialConsumer Reports, TechRadar, Tom's Guide and The Verge in tech; MotorTrend, US News and Edmunds in autos. These are the peers to out-authority in your niche.
- forbes.comEditorial / ListicleRecurs across verticals and is rising on ChatGPT per Semrush. The kind of large-domain listicle AI trusts.
linkedin.comProfessional networkA top-five source on Google AI Mode and Perplexity for B2B topics, and where creator credibility is read.- Vertical reference authoritiesReferenceNIH for health, Investopedia and NerdWallet for finance. The niche authority you have to beat to be the cited source.
These source patterns trace to our four-engine AI Citation Study; re-check them as the category moves.
publishers and creators-specific moves
- Own a defined niche, not a broad oneAI cites the authority in a specific vertical. Pick a niche narrow enough to be the best source in, then be more current and more structured than the reviewers already cited there.
- Structure every piece to be liftedLead with the answer, use headings as questions, add schema, and keep listicles dated and maintained. A model quotes the passage it can extract cleanly.
- Live on the platform layer AI readsGet your work discussed on Reddit, present on YouTube, and anchored by a Wikipedia and LinkedIn entity, because that layer recurs in every engine's citations.
- Measure your citation share per engine, over timeWhich publishers win swings by engine and by month, so track your share in ChatGPT, Perplexity and Google separately rather than chasing one fixed order.
This is the one segment where the citation is not a means to a sale, it is the product. Prefer (our product) measures your visibility on those terms, tracking which publishers each engine cites for your niche’s prompts on ChatGPT, Gemini, Perplexity, Google AI Overviews and AI Mode. The sections above frame it that way: the prompts a publisher competes on, which engines answer them, and the publishers AI already reads, listed as your competitive set rather than a channel to court. The threat underneath it is the collapsing click, and the response is not to fight for the click but to win the citation and the authority that comes with it. The plays follow, in the order a publisher or creator starting from low visibility should run them.
The plays that win publisher and creator citations#
Be the current authority in a defined niche
Choose a niche narrow enough to be the best, most current source in, and out-authority the reviewers AI already cites there, because AI names the vertical authority.
Across verticals, AI cites a small set of category-authority publishers plus the platform layer. You cannot out-authority Wikipedia or Consumer Reports everywhere, but you can be the most current, most specific source in a defined niche, which is what wins the recommendation and explainer prompts that used to send you search traffic. (Semrush most-cited domains study, checked 2026-09-05)
- Pick your niche and list the recommendation, explainer and comparison prompts readers ask in it.
- Run those prompts and record which publishers AI cites, that is your competitive set.
- Audit your coverage against theirs: what is more current, more specific or better-tested on your site.
- Fill the gaps and keep the cornerstone pieces updated, because freshness is part of why AI picks a source.
Done when: You are the most current, most specific source for a defined set of prompts in your niche.
Verify it worked: Ask your niche's recommendation and explainer prompts in ChatGPT and Perplexity and check whether you are cited alongside or above the incumbents.
Common failure mode: Chasing broad, high-competition topics where entrenched authorities and Wikipedia already own the citation. You get summarized, not cited.
Structure every piece so AI can lift it
Make every piece extractable: lead with the answer, use question-shaped headings, add schema, and keep listicles dated, so a model lifts your passage instead of a rival's.
AI cites the source it can extract cleanly. A piece that buries its answer under a long intro, or hides its verdict in prose, gives the model nothing liftable, so it quotes a competitor whose structure is clearer, even when your reporting is better.
- Open each piece with a self-contained answer or verdict a model can quote without context.
- Write headings as the questions readers actually ask, and answer in the first sentence under each.
- Add Article, FAQPage and, for how-tos, HowTo schema; mark up author and publish or update dates.
- Date and maintain your roundups and 'best of' lists, and state the criteria you used.
Done when: Your cornerstone pieces open with a liftable answer, carry valid schema, and show clear, current dates.
Verify it worked: Ask the question a piece answers and check whether your wording, or a rival's, is what the engine quotes.
Common failure mode: Answer-last articles and undated listicles. AI cannot tell if they are current or lift a clean claim, so it prefers a structured competitor.
Get onto the platform layer AI reads
Get your work discussed on Reddit, present on YouTube, and anchored by a Wikipedia and LinkedIn entity, because that platform layer recurs in every engine's citations.
Reddit was the single most-cited source in our study (30.3% on ChatGPT) and Semrush's, and YouTube tops Google's citations. AI often finds and corroborates your work through this layer, so a piece that is never discussed on it is far less likely to be cited, however good it is. (Prefer Reddit study, checked 2026-08-19)
- Take part honestly in the subreddits and communities where your niche is discussed; share genuinely useful work.
- Repurpose your reporting into video on YouTube, the top source on Google's AI answers.
- Make sure your publication and key creators have accurate, well-sourced Wikipedia entities where they qualify.
- Keep author LinkedIn profiles current, since LinkedIn is a top source on Google AI Mode and Perplexity for B2B.
Done when: Your work is genuinely discussed on the communities, present on video, and anchored by accurate entities.
Verify it worked: Ask a niche prompt and check whether the Reddit threads, videos and reference pages cited point back to your work.
Common failure mode: Dropping links into communities to game them. It gets removed and discounted, and a planted mention corroborates nothing.
Measure your citation share per engine, over time
Track your citation share in ChatGPT, Perplexity and Google separately, over time, because the winning publisher mix differs by engine and shifts month to month.
Semrush documented Reddit swinging from close to 60% of ChatGPT responses to around 10% in six weeks, and Wikipedia leading ChatGPT while nearly absent on Google AI Mode. A single-engine or single-snapshot view will mislead you, so the honest metric is your own share per engine over time, not last quarter's order. (Semrush most-cited domains study, checked 2026-09-05)
- Define a set of the prompts you want to be cited for in your niche.
- Run them on a schedule across ChatGPT, Perplexity and Google, and log whether you are cited on each.
- Track your share per engine over time, not a single blended number.
- Feed the gaps back into the plays above, doubling down where you are close and rethinking where you are absent.
Done when: You have a per-engine citation-share trend for your target prompts that you review on a cadence.
Verify it worked: Re-run your prompt set and confirm the trend moves as you ship the plays above.
Common failure mode: Judging yourself on one engine or one snapshot. The mix moves, so a single reading tells you little about where you actually stand.
How this fits your existing SEO and audience strategy#
None of this throws out your SEO or your audience playbook. The clear structure, authority and freshness that helped you rank are the same traits AI rewards, so a strong publisher has real groundwork to build on. What changes is the goal: SEO optimized for the click, AEO optimizes for the citation, because the click is disappearing. You measure it differently too, citation share per engine over time rather than position, and you use that share for brand authority, subscriptions and direct visits rather than raw pageviews.
Start with what AEO is for the full method, or see the AEO for B2B SaaS playbook for how a content-heavy business wins the same citations. You can also browse every AEO-by-business-model playbook.
A worked example
A review authority AI repeatedly cites, the pattern of a structured, trusted vertical source.
Cited across our tech and automotive pulls; reliability data AI quotes as fact.
A top-five domain in our Reddit study's ChatGPT set, a category-authority reviewer that wins recommendation prompts.
A finance-vertical authority AI reaches for on money questions.
A definitional-authority publisher that competes with Wikipedia in finance.
Named brands are public, illustrative examples of the category, not customers or endorsements.
Sources
People also ask
- How do publishers get cited by ChatGPT?
- How do creators get cited by AI as a source?
- Which publishers does AI cite most?
- Do AI summaries reduce publisher traffic?
- How is AEO different from SEO for a media brand?