Many business owners assume there is one “AI search” to optimise for. They hear that a brand appears in ChatGPT and expect it to show up the same way in Perplexity and Google's AI Overviews. “If I get cited by one, won't the others follow?”
The data says no, at least not reliably. The sources each AI assistant leans on are surprisingly different. A site that ChatGPT mentions all the time may barely appear in Perplexity, and the reverse.
Let me explain how little the top sources overlap, why each assistant reaches for different websites, and how to plan for each platform without tripling your workload. I will finish with a simple way to track and report visibility per platform.
How much do the top sources overlap across AI assistants?
The clearest data comes from Patrick Stox at Ahrefs. In 86% of Top Mentioned Sources Are Not Shared Across ChatGPT, Perplexity, and AI Overviews (June 12, 2025), he compared the 50 most-mentioned websites on each platform, using June 2025 data from about 76.7 million AI Overviews, 957,000 ChatGPT prompts and 953,500 Perplexity prompts.
Only 7 websites appeared in the top 50 on all three platforms. That is 14%. The other 86% of each top list was not shared across all three.
Other studies, using different methods and dates, find a bit more overlap, but never a lot:
| Study | Platforms compared | Overlap finding |
|---|---|---|
| Ahrefs (June 2025) | ChatGPT, Perplexity, AI Overviews | 7 of top 50 shared by all three (14%) |
| Similarweb (April 2026, U.S. data Jan–Feb 2026) | ChatGPT, Google AI Mode | 13 of top 20 domains shared; 7 unique to each |
| Semrush (November 2025, data July–October 2025) | ChatGPT search, AI Mode, Perplexity | Reddit and LinkedIn in the top five on all three |
So the giant, general platforms (Wikipedia, Reddit, YouTube, LinkedIn) tend to be shared. Below them, the lists split quickly. That “below the giants” zone is exactly where most businesses compete.
“Good content is good content, and top websites are top websites.”
— Patrick Stox, Ahrefs, 86% of Top Mentioned Sources Are Not Shared Across ChatGPT, Perplexity, and AI Overviews
That quote is a fair balance. Strong sites do well everywhere. But each platform then adds its own preferences on top.
The domains each assistant mentions most
Here is what the studies show about each platform's favourite sources. Different studies measure different things (mentions vs citations, share of all citations vs share of responses), so compare within a row, not across rows.
| Platform | Top sources reported | Source and date |
|---|---|---|
| ChatGPT | Wikipedia 16.3% of mentions; Reuters, Apple News (AP) and AS.com in the top 10 at about 2.6%–4% | Ahrefs, June 2025 data |
| ChatGPT | wikipedia.org 13.15%, reddit.com 11.97%, openai.com 6.21%, walmart.com 2.90%, youtube.com 2.67% | Similarweb, U.S., Jan–Feb 2026 |
| Perplexity | YouTube 16.1%, Wikipedia 12.5%; niche health sites such as tuasaude and alodokter also rank | Ahrefs, June 2025 data |
| Perplexity | Reddit, LinkedIn, NIH, Microsoft, Google; Wikipedia in only about 0.8% of responses | Semrush, July–Oct 2025 |
| Google AI Overviews | YouTube 9.5%, Wikipedia 8.4%, Reddit 7.4%, Quora 3.6%; Mayo Clinic and Cleveland Clinic also appear | Ahrefs, June 2025 data |
| Google AI Mode | fandom.com 7.16%, wikipedia.org 5.21%, youtube.com 4.91%, reddit.com 4.19%, google.com 2.85% | Similarweb, U.S., Jan–Feb 2026 |
I have written a separate deep dive for each platform, with what the lists mean for your industry: the sites ChatGPT cites most, the sites Perplexity cites most, the sites Google AI Overviews cite most and the sites Google AI Mode cites most.
Why each assistant reaches for different sources
There is no single reason. Four things combine.
1. They search different indexes
Google's AI features sit on Google's own search index. Google's AI features documentation (updated December 2025) says a page must be indexed and eligible for a snippet to be a supporting link, and that both AI Overviews and AI Mode may issue “multiple related searches across subtopics and data sources.” So Google's AI picks from what Google already ranks.
ChatGPT and Perplexity run their own search setups, with their own crawlers and partners. If you want to know where ChatGPT gets its results, see my guide on which search index ChatGPT uses.
2. Partnerships and licensing
Ahrefs links ChatGPT's heavy use of news, entertainment and sports publishers partly to licensing deals. Semrush found Google AI Mode leaning on Google-owned properties like YouTube and Google's own sites. Business relationships shape which sources are easy for each assistant to use.
3. Product design
Ahrefs summarised how each system handles citations. Here is my plain-language version:
| System | How it finds and cites sources (per Ahrefs) | What that means for you |
|---|---|---|
| Perplexity | Real-time web search, then ranks, quotes and cites the best match | Being findable and quotable right now matters most |
| ChatGPT with browsing | Searches, filtered toward trusted or licensed domains | Trusted publishers and reference sites carry extra weight |
| ChatGPT without search | No citations unless prebuilt; may make up sources | What the model learned about you in training matters |
| Google AI Overviews | Draws from the search index; curated citation list | Normal Google rankings are the entry ticket |
| Bing Copilot | Uses Bing search, highlights snippets, includes citations | Bing indexing and rankings matter |
4. Who blocks whom
Some sites block AI crawlers, which removes them from that assistant's pool. Ahrefs' AI bot block rate study (May 21, 2025) found about 5.6% to 5.9% of around 140 million websites blocked GPTBot, ClaudeBot, PerplexityBot or Google-Extended through robots.txt. Each site blocks a different mix, so each assistant sees a slightly different web. It is worth checking you are not blocking a crawler by accident. My guide on verifying AI crawlers explains how.
What each platform seems to prefer
Pulling the studies together, these are the patterns I would plan around. They are tendencies, not rules.
ChatGPT leans on reference and news sources:
- Wikipedia is consistently its top or near-top source.
- News publishers such as Reuters and AP feature strongly (Ahrefs, 2025), and Forbes roughly doubled its ChatGPT citation rate after September 2025 (Semrush).
- Ahrefs found little health or medical content among its top sources.
Perplexity leans on video, reference and a wider international mix:
- YouTube and Wikipedia led its mentions in Ahrefs' June 2025 data.
- It included many local and regional brands, and few social sites or traditional Western media.
- Ahrefs' link metrics study (June 2025) also found Perplexity's mentions tracked link strength more than the other two did, though still weakly.
“Perplexity doesn’t have social websites or traditional Western media.”
— Patrick Stox, Ahrefs, 86% of Top Mentioned Sources Are Not Shared Across ChatGPT, Perplexity, and AI Overviews
Google AI Overviews and AI Mode lean on community and Google's own properties:
- YouTube, Reddit and Quora rank highly (Ahrefs, 2025), along with authoritative health sites like Mayo Clinic.
- AI Mode showed a clear preference for Google-owned properties (Semrush, 2025).
- AI Overviews track brand fame closely. In another Ahrefs study (July 2025), web mentions correlated 0.65 with AI Overview visibility, against 0.15 for ChatGPT.
Sources shift fast, so plan for drift
Even within one platform, the favourite sources move. Semrush tracked weekly snapshots from July to October 2025. On ChatGPT, Reddit fell from close to 60% of responses in early August to around 10% by mid-September, and Wikipedia fell from about 55% to under 20%. LinkedIn rose steadily across platforms. Semrush noted that the data shows what changed, not why.
The lesson: do not build a plan around one source on one platform. Spread your presence so a single change cannot wipe you out.
Choosing which AI platforms to prioritise
You probably cannot do everything at once. Here is how I would decide.
Start with where the traffic and the audience are. SE Ranking's June 2026 study of 101,574 sites found ChatGPT sent 74.78% of AI referral traffic in early 2026, followed by Gemini (11.56%), Perplexity (7.23%), Copilot (3.51%) and Claude (2.62%). Google AI Overviews send their clicks as normal organic search, so they do not appear in those numbers, but they reach anyone who uses Google.
| If your business is… | Prioritise first | Then | Why |
|---|---|---|---|
| Local service (plumber, dentist, salon) | Google AI Overviews | ChatGPT | Local searches still start on Google; AI Overviews sit inside them |
| B2B software or services | ChatGPT | Perplexity, AI Mode | Most AI referral traffic comes from ChatGPT; research-heavy buyers also use Perplexity |
| Ecommerce | Google AI Overviews and AI Mode | ChatGPT | Shopping research still runs through Google; ChatGPT cites retailers like Walmart |
| Health or finance publisher | Google AI Overviews | Perplexity | AI Overviews favour authoritative health and finance sites; Perplexity cites niche health sites |
| News or entertainment | ChatGPT | AI Mode | ChatGPT leans on media publishers; AI Mode cites fan and entertainment sites |
This matrix is my own judgment based on the studies above. Your best check is to ask your customers, or look at which assistants already send you visits in GA4.
Shared groundwork vs platform-specific tactics
The good news: about half the work helps everywhere. Let me split it.
Shared groundwork that helps on every platform:
- Clear, specific pages that answer real questions. Similarweb found AI rarely cites homepages and prefers deeper pages.
- Strong Google and Bing rankings, since several assistants draw on search results.
- Crawler access for every AI search bot you want to appear in.
- Presence on the shared giants: an accurate Wikipedia or Wikidata entry if you qualify, active YouTube content, and honest participation on Reddit and LinkedIn.
Platform-specific tactics:
- ChatGPT: press coverage in trusted news and trade publications; being on the reference pages it already cites.
- Perplexity: fresh, quotable pages with clear facts; YouTube videos; regional and niche sites in your market. See my guide to Perplexity SEO.
- Google AI Overviews and AI Mode: classic SEO plus YouTube and community presence. My guide to YouTube SEO and Reddit for AI visibility cover both.
A worked example: one plan, three platforms
This is a hypothetical example. A B2B payroll software company tracks 40 buyer questions across three platforms (120 answers). It is named in 22 ChatGPT answers (55%), 9 Perplexity answers (23%) and 6 AI Overview answers (15%).
An averaged score of 31% would hide the real story. Broken out, the plan becomes obvious:
- Protect ChatGPT by keeping its press mentions and comparison-page listings current.
- Fix AI Overviews first, because it is weakest there and the reach is largest. The cited pages are mostly Google-ranked guides and Reddit threads, so the company improves its own guides and joins those discussions honestly.
- Grow Perplexity with short explainer videos on YouTube and updated, fact-dense help pages.
- Re-measure monthly and move effort to whichever platform is lagging.
Tracking AI visibility platform by platform
Because sources differ, you need to track each platform separately. Here is the routine I suggest.
- Fix one question set. 30 to 50 real customer questions, unchanged from month to month.
- Run it on each platform. At minimum ChatGPT, Perplexity and Google (AI Overviews and AI Mode). Use a fresh session, and run important questions twice because answers vary.
- Record three things per answer: whether your brand is named, whether your page is cited, and which third-party domains are cited.
- Build a source list per platform. The cited domains become your outreach and content targets for that platform.
- Add traffic data. Pull AI referral visits by source from analytics so visibility connects to visits.
Doing this by hand is fine for 30 questions. Beyond that, a tool helps. I compare the options in AI visibility tools compared. If you are new to the platforms themselves, start with my overview of AI search engines.
Reporting AI visibility without averaging the platforms together
The biggest reporting mistake I see is one blended “AI visibility score.” Since each platform cites different sources, a blended number can stay flat while you gain on one platform and lose on another.
A better report shows, for each platform: share of answers naming you, share citing your pages, your top three rivals' shares, the top cited third-party sources and any change from last month. Then add a short note on what you will do next on each one. For a fuller method, see my guide to AI share of voice.
Questions marketers ask about AI assistants citing different sources
Why does ChatGPT cite different websites from Google AI Overviews?
They pull from different places. Google's AI Overviews use Google's own search index, while ChatGPT combines its training with its own search setup and leans toward news publishers and Wikipedia. Different sources go in, so different sources come out.
How much do AI assistants' top sources overlap?
Very little at the very top. Ahrefs found only 7 of the 50 most-mentioned websites were shared by ChatGPT, Perplexity and AI Overviews in June 2025, which is 14%. Similarweb found more overlap between ChatGPT and Google AI Mode, with 13 of each top 20 shared in early 2026.
Which AI platform should a small business focus on first?
Usually ChatGPT for referral traffic, because it sends most AI visits, and Google AI Overviews for reach, because they appear inside normal Google searches. Then check which platform your own customers use before adding others.
Is there any source that works across all AI assistants?
A few come close. Semrush found Reddit and LinkedIn among the top five cited domains on ChatGPT, AI Mode and Perplexity in 2025, and Wikipedia and YouTube rank highly on several platforms. Your own well-built pages and good search rankings also help everywhere.
How often do the most-cited sources change?
Often enough to matter. Semrush saw Reddit fall from close to 60% of ChatGPT responses to around 10% within weeks in 2025. I suggest reviewing your target source list every quarter.
Planning for several AI assistants without tripling the work
So, why does each AI assistant cite different sources? Because each one searches a different index, has different partners, is built differently and sees a slightly different web. Ahrefs found only 14% of the top 50 sources shared across ChatGPT, Perplexity and AI Overviews. The giants overlap; the rest does not.
The practical answer is not three separate strategies. It is one shared foundation, a few platform-specific moves, and tracking that keeps each platform in its own column.
If you would like us to map where your brand appears on each assistant and build the plan, see our SEO services or get a free quote.
Further reading on AI platform source overlap
- AI Visibility Tools Compared: Options for Tracking Brand Mentions in AI
- How ChatGPT Turns One Prompt Into Many Web Searches
- Do LLMs Flatten Original Ideas? Keeping Your Brand's Point of View in AI Answers
Sources for this AI platform source overlap guide
- Ahrefs: 86% of Top Mentioned Sources Are Not Shared Across ChatGPT, Perplexity, and AI Overviews (June 12, 2025)
- Ahrefs: The 10 most mentioned domains for ChatGPT, Perplexity and AI Overviews (June 11, 2025)
- Ahrefs: Domain-Level Link Metrics May Not Be Good Predictors of AI Search Mentions (June 25, 2025)
- Ahrefs: Google Seems More Biased Towards Big Brands Than ChatGPT and Perplexity (July 7, 2025)
- Ahrefs: AI bot block rates (May 21, 2025)
- Semrush: The most-cited domains in AI, a 3-month study (November 10, 2025)
- Similarweb: The most cited domains by LLMs (April 14, 2026)
- Google Search Central: AI features and your website (updated December 10, 2025)
- SE Ranking: AI traffic grew 16x from 2024 to 2026 (June 18, 2026)
