Briefing

Which websites AI engines cite most — and how to earn a citation

The domains AI cites most are the ones you'd expect — Reddit, Wikipedia, YouTube, the big directories and news sites — and no local business is going to displace them. The winnable game is smaller and more useful: being the source an engine trusts when the question is about your category, in your city.

Updated August 29, 2026Reading time 5 min

The most-cited lists all point at the same handful of sites

If you read the analyses that tried to rank which domains AI engines cite most, the striking thing is how boring the top of the list is. Across AI Overviews, AI Mode, Gemini, and Perplexity, mid-2026 studies — several of them vendor-published by SEO-tool companies like Ahrefs, which is reason enough to treat the exact figures skeptically — kept surfacing the same names at the top: Reddit, Wikipedia, YouTube, the large directories, and the major news outlets. Different methodologies, different engines, roughly the same leaderboard. One recurring finding is worth flagging on its own: AI Mode appears to lean on a relatively small, trusted set of sources rather than casting a wide net. The engines are not citing the whole web evenly. They’re citing a short list of places they already trust, over and over.

Why a vendor’s leaderboard flatters itself: most “most-cited domains” studies count raw citation volume across every kind of query — recipes, medical questions, celebrity trivia, product reviews — so of course Reddit and Wikipedia win. That tells you what AI cites in general. It tells you almost nothing about what AI cites when someone asks for a plumber in your city. Those are different questions, and only one of them is your business.

Why the global leaderboard is the wrong target

Read those lists as context, not as a to-do. You are not going to out-cite Wikipedia, and you don’t want to — the effort it would take to become one of the web’s most-cited domains has nothing to do with running a law firm or an HVAC company. Worse, the framing quietly sets you up to lose. If the goal is “get onto the list AI cites most,” the answer is permanently no, and you’ll walk away concluding AI visibility is hopeless. It isn’t. You’ve just been handed the wrong scoreboard.

The giants win the global count because they answer the whole internet’s questions. Your customers aren’t asking the whole internet’s questions. They’re asking a narrow one — who’s the best in this category near me — and when they do, the engine reaches for a different, much shorter list.

The list that matters is the one for your category, in your city

Here’s the part the leaderboards obscure. When someone asks an engine a local, buying-intent question, it doesn’t rank the whole web. It pulls from the sources it trusts for that specific question: the directory that dominates your vertical, your Google Business Profile and its reviews, the review platforms the engine reads, and the handful of pages — yours or a competitor’s — that answer the question in quotable form. That shortlist is winnable. It’s a short, knowable list rather than the whole web, and several of the entries are things you control or can influence directly.

So the real question isn’t “how do I get on the most-cited list?” It’s “when a customer asks about my category in my city, which sources does the engine cite — and am I one of them?” That one you can answer, and then act on.

What our own archive says, per industry

The lists above come from other people’s analyses. Here is ours, computed across our own measurement panels in real local markets, on six engines.

The sample: 3,194 runs across 14 panels, of which 1,647 cite anything at all — two of the six engines return no sources by design, so the figures below are the share of citing runs. Averaging in engines that never cite anything tells you about our engine mix rather than about your industry. Results are split by sub-industry, which turns out to matter.

industry runs industry directories Yelp official registries
family law 1,053 65.8% 18.7% 17.2%
med spa 540 22.4% 25.4% 5.3%
personal injury 531 47.1% 1.7% 20.3%
cannabis 446 70.7% 19.6% 24.9%
elder care 208 61.6% 4.1% 27.4%
plastic surgery 208 55.5% 24.2% 1.6%
addiction treatment 208 68.4% 3.8% 43.0%

The unit is the sub-industry, and that is the most useful thing on this page. Med spa and plastic surgery sit next to each other in any normal taxonomy, and they have opposite source mixes: a med spa’s customers are found through Yelp, a plastic surgeon’s through directories. Roll them together into “aesthetics” and the finding disappears — which is exactly what our first pass did, and it briefly produced the wrong answer. If someone hands you an off-site strategy for your category, ask whether they measured your niche.

Med spa is the only one of the seven where Yelp beats directories. That is the counterintuitive result on this page and it survived re-measurement.

Addiction treatment has the highest registry share of any industry we measure — 43.0%, roughly 1.6x elder care and 8x family law. Registries matter more there than anywhere else, though directories still lead them, 68.4% to 43.0%.

For law, the directory layer breaks down in a clear order — and the order is the finding, not the number. Super Lawyers, then Justia, then Avvo, then lawyers.com. That top four came back identical on re-measurement while the absolute percentages nearly doubled, purely because the denominator changed. Treat the ranking as the instruction and ignore the decimals: Super Lawyers 13.4%, Justia 10.3%, Avvo 10.2%, lawyers.com 6.5%, Expertise.com 6.3%, FindLaw 4.7%, Martindale 3.7%, LawInfo 3.0%.

Reddit is the most-cited third party we cannot place you on — 6.8% of all runs, across 13 panels. It is also the one nobody can place themselves into, and anyone offering to is offering to write fake posts. That is the honest ceiling on this kind of work.

One limit worth stating plainly: this measures what the engines cite, not proof that adding a business to those sources changes what they cite. It tells you where to look and in what order. It is not a causal claim, and we will say so until we have run the controlled version.

How to become one of the cited sources

The work sorts into three moves, and none of them is glamorous.

  • Be present, complete, and consistent on the directories the engines actually read for your category. Different engines trust different directories, and the ones that matter for lawyers aren’t the ones that matter for restaurants. A thin or missing profile on the directory an engine leans on is a citation you’re handing to a competitor.
  • Earn genuine reviews that say specific things. Engines increasingly read review text, not just the star average, and quote the specifics back. Reviews that mention the outcome, the price, or the responsiveness become the phrases an engine reaches for when a prospect asks about exactly those things.
  • Publish the quotable page. A lot of local questions — cost, process, “do I need this” — go unanswered by any business in the market, so the engine quotes whoever bothered to write it down clearly. A plain, specific page that answers the real question is often the cheapest citation you’ll ever earn.

None of this puts you on a national leaderboard, and no honest process can promise a given engine will cite you. But all of it makes you the kind of source an engine reaches for when it matters — in your category, in your city.

Start with your own list, not the global one

The global most-cited domains are worth knowing about and useless as a plan. The useful version is yours: for your category and your city, across the six engines your customers use, which sources get cited, and where you stand in them. That per-market picture is what Nameworthy puts on paper, with every source on file so you can check the work — and then the fix list writes itself, directory by directory, review by review, page by page.

If you want the whole approach in order, read how to make your business show up in AI. If you’d rather just see who AI cites in your market right now, start here.

Common questions

Which websites do AI engines cite most?
According to vendor-published analyses in mid-2026, the top of every list looks similar across AI Overviews, AI Mode, Gemini, and Perplexity: Reddit, Wikipedia, YouTube, the large directories, and major news outlets. One recurring finding is that AI Mode appears to lean on a relatively small, trusted set of sources rather than the open web. Treat the exact figures skeptically, since much of the ranking work is published by companies selling SEO tools.
Can my business get onto the most-cited list?
Not the global one, and you shouldn't spend effort trying. Displacing Wikipedia or Reddit in overall citation volume has nothing to do with running your business. The winnable game is the local one: being cited when someone asks an engine about your category in your city, which comes from the directories, reviews, and quotable pages the engines trust for that specific question.

The report is free

Find out who AI cites in your market.

Your real customer questions across the major engines, every source on file.

Get your AI visibility report