YouTube in AI Answers: Google Cites Your Videos, ChatGPT Barely Notices Them

YouTube has been the most-cited source across AI answer engines since spring 2026, ahead of Reddit. That headline ranking hides a wide gap between engines. On some topics, Google's AI products cite a video in more than one answer in three, while ChatGPT cites almost none. If you're a small business or an agency weighing whether video is worth the investment, you need to make that call engine by engine, not for "AI" as a whole.

This article draws on three recent studies, including one from the agency Foundation, last updated October 24, 2026, and built on tens of thousands of YouTube citations. It covers where YouTube stands in each engine, what the most-cited videos have in common, and how Google links to a specific moment rather than to the whole video. It closes with a practical starting point sized for a small-business budget.

Key takeaways

• According to Meltwater, YouTube remains the most-cited source across AI engines combined, despite its citations falling by more than half in September 2026.
• Google's AI products account for most of that lead, while ChatGPT almost never cites YouTube.
• In Foundation's study, small channels are well represented among the most-cited videos, whose titles often mirror a dated comparison question.
• Foundation also found that nearly half of Google's YouTube citations point to a specific timestamp, which makes chapters worth adding.
• None of these studies breaks out results by language or market, which argues for testing a handful of videos rather than overhauling your strategy.

How much of the AI answer landscape does YouTube own?

YouTube is now the single most-cited source across AI answer engines taken together, ahead of Reddit. That position, however, owes far more to Google's products than to any other assistant.

Meltwater's AI Search Visibility Report, published July 3, 2026, analyzed more than 8 million citations collected in May 2026 across eight AI models. The set includes ChatGPT, Gemini, Perplexity, Claude and both of Google's search products. YouTube came out on top with 188,863 citations for the month. Reddit, in second place, logged 111,860 according to the same report.

The authors also note a 56.4% jump in YouTube citations in a single month. They call YouTube optimization "a core GEO lever," GEO being generative engine optimization, the work of earning citations in AI answers. That momentum didn't hold: YouTube citations fell by more than half in September 2026, to 92,855, according to Meltwater's report published October 15, 2026. YouTube still ranks first, and Google's AI products still treat it as their leading source. The report offers no explanation for the drop.

The per-engine breakdown for May, below, tells a more nuanced story.

Source: Meltwater, AI Search Visibility Report, June 2026 edition (May data).

Google's three products account for the lion's share. Perplexity is a steady but much smaller contributor. ChatGPT, by contrast, shows no meaningful YouTube citation volume, even though it sends by far the largest share of traffic from AI assistants. This report gives no figure for Claude. The same video investment, in other words, will pay off very differently depending on which engine your customers use.

Google cites YouTube; ChatGPT mostly ignores it

Another study confirms the gap on a narrower scope: on software buying questions, a Google answer is about 17 times more likely to cite YouTube than a ChatGPT answer.

Every week from June 29 to September 21, 2026, Foundation ran the same B2B software buying questions through ChatGPT, Google AI Mode and Google AI Overviews. The sample covered 5,000 questions spread across eighty software categories. Over the full period, 83,835 citations pointed to a YouTube video, according to Foundation's study, last updated October 24, 2026.

Source: Foundation, B2B Citation Index, October 2026.

Google's products generated 98.2% of all the YouTube citations the study recorded. Even on reviews and validation questions, its strongest category, ChatGPT cited a video in just 4.1% of answers. Query type matters on Google too: 46.5% of answers to "what's the best tool for…" questions cited a video. That share drops to 19.1% for pricing questions.

The study doesn't explain the gap between engines, so any interpretation is a hypothesis for now. The most plausible one is ownership: Google owns YouTube and has every video's title, description and transcript in its own index. ChatGPT draws on a different pool of sources and appears to favor written pages. In May 2026, its two most-cited sources were Wikipedia and the US National Institutes of Health (NIH) website, according to Meltwater. For your strategy, that means a video works first for your visibility in Google, while your presence in ChatGPT still depends on your pages and on how often the press and independent sites mention your brand.

What the most-cited videos have in common

The videos Google cites most aren't the ones with the biggest audiences. They stand apart on title, length, structure and publication date.

Among the 150 most-cited videos, Foundation found almost no relationship between view count and citation frequency, with a near-zero rank correlation of −0.08. Likewise, 42% of the most-cited videos came from channels with fewer than ten thousand subscribers. The single most-cited video in the entire study had only 3,671 views at the time of measurement. As the authors put it, "audience size barely mattered."

The table below compares the 100 to 150 most-cited videos with those cited only once.

Source: Foundation, B2B Citation Index, October 2026.

The most-cited videos tend to put the buyer's question right in the title, often as a dated comparison. Product demos, useful as they are during a trial, get picked up far less often in AI answers. Even so, seven of the thirty most-cited videos came from software vendors' own channels, and nearly all of them compared the whole category, competitors included. Fresh content dominates the top of the list as well: 69% of the most-cited videos were less than a year old, which mirrors what we see with written pages and citation decay.

How does Google find the key moments in a video?

Google often cites a specific moment in a video rather than the video as a whole. Chapters are the most direct way to point it to those moments.

In Foundation's study, 48.4% of Google's YouTube citations pointed to a specific timestamp. The median cited moment sat at 2 minutes 28 seconds, meaning half of all cited segments appeared before that mark. Roughly 71% of them fell within the first five minutes. ChatGPT, by contrast, produced zero timestamped citations over the entire period. "Google cites the moment," the authors write.

This builds on a mechanism Google has documented for years in its video results, known as key moments. Google's developer documentation says that for videos hosted on YouTube, timestamps and labels in the description let you specify them. For videos hosted on your own site, Clip markup within structured data of type VideoObject lets you set the start, end and label of each segment. Google also recommends, where it makes sense for your business, giving each video a dedicated watch page where the video is the main content.

On the YouTube side, the official help page on chapters sets three rules for manual chapters. The first timestamp must be 00:00, the list needs at least three timestamps in ascending order, and each chapter must last at least ten seconds. A well-named chapter then works much like a subheading in an article. Foundation recommends naming chapters after buyers' questions and delivering the verdict early, which matches where cited segments cluster.

Brand mentions in YouTube videos

Brands that get named often in YouTube videos, including other creators' videos, also tend to be the most visible in AI answers.

Ahrefs' study published December 12, 2025 compared a range of metrics against the visibility of 75,000 brands in ChatGPT, AI Mode and AI Overviews. Ahrefs defines a "YouTube mention" as any appearance of a brand name in a video's title, transcript or description. That factor showed the strongest correlation of all those tested, at about 0.737 on a scale from −1 to 1. Branded web mentions and backlinks ranked below it.

Ahrefs is careful to point out that correlation isn't causation, and that improving these metrics won't automatically lift your AI visibility. It's possible that well-known brands simply get mentioned more in videos and in AI answers alike, with neither driving the other. For a smaller company, the practical takeaway is to identify the creators already comparing products in your space. You can then give them something worth covering, such as access to your product or original data.

What these studies don't tell you

All three studies describe trends and correlations, and none states the language of its prompts. They point to a test worth running, not a settled answer for every market.

Foundation's study covers B2B software buying queries only, a category where video comparisons are plentiful. Its authors stress that their findings describe cited videos and don't prove what causes a citation. The share of AI Mode answers citing YouTube also swung between 17% and 56% across weekly runs during the study. Meltwater counts raw citations across a fixed set of prompts, so engines that cite many sources per answer naturally carry more weight in its ranking. None of the three studies breaks out results for non-English queries.

These caveats don't undo the core finding, which holds across studies using different methods: Google leans heavily on YouTube, while ChatGPT barely uses it. They do mean you should check, in your own sector and on your own questions, whether videos appear in AI answers before committing a production budget. A one-off check won't do either, given how widely these numbers swing over time.

Where to start on a small-business budget

Before producing new videos, confirm that Google already cites videos for your market's questions, then rework the videos you already have.

Start by listing twenty or so buying questions specific to your sector and running them through Google, in both AI Overviews and AI Mode, then ChatGPT. If no videos appear in Google's answers, video probably isn't your priority right now. If competitors' or creators' videos do appear, note their titles, runtimes and the segments being cited, because they show you the format the engine expects. A citation tracking tool lets you repeat that check weekly in AI Overviews and ChatGPT without losing half a day to it.

Next, turn to your existing videos, which take little effort to put to better use. Rewrite their titles to reflect the buyer's question, adding the year when the topic changes over time. Split them into chapters named as questions, and make sure the answer or verdict lands in the first few minutes. Finally, proofread YouTube's automatic transcript, where a misspelled brand name easily goes unnoticed. It's one of the three places, along with the title and description, where Ahrefs counted brand mentions.

If you do decide to produce, favor an honest five-to-fifteen-minute comparison of your category over a demo of your product alone. Plan to refresh it once a year, since most of the top-cited videos are recent. Publish the video on a page of your own site as well, with structured data and a text transcript. That page could help with ChatGPT, which seems far more inclined to read written pages than videos. Our AI visibility audit checks that the page stays readable for AI crawlers.

Our recommendation is to treat YouTube as a Google-specific channel and test it on a few videos for at least a month before scaling up. Measure citations before and after each change, using the method in our guide on proving your GEO is working. Your visibility in ChatGPT, for its part, will keep depending on your written pages and on what other sites say about you.