A B2B study says 88 percent of AI-quoted sources are brand-controlled pages: Reveneau

Why it mattersIf the split holds outside this sample, a B2B team chasing AI visibility should spend its next hour on its own product, comparison and how-to pages, not on Reddit threads or YouTube uploads.

Ten Speed, a B2B marketing agency, published a study on 3 September that measured which pages ChatGPT, Perplexity, Claude, and Gemini quote when a buyer asks a real evaluation-stage question. Search Engine Journal reported it the same day, along with a fact-check exchange between the outlet and the agency about how the numbers were produced.

What the sample was

The study, written by Nelson Brassell of Ten Speed, tracked 7,387 citations across 170 prompts. Ten Speed drew the prompts from real evaluation-stage buyer language across its own clients in fintech, physical security, hospitality, and IT automation. The four engines were monitored through Peec AI. Search Engine Journal notes that the sample is one agency's client base rather than a random draw, which limits how far the split can be generalised.

What the split looks like

Ten Speed's main finding is that pages the brand owns make up most of the citations at this stage of the buying process. Product pages accounted for 24.1 percent of citations. Articles, including blog posts, news and PR, took 17.4 percent. Comparison guides and listicles each took about 13 percent. How-to guides came in at 8.9 percent and homepages at 7.8 percent. Together, brand-controllable content reached 88.3 percent of the total.

Community and social sources came in far below what many AI-visibility discussions this year have assumed. Reddit, YouTube and other forums together accounted for 4.2 percent of citations. Directory profiles on sites like G2 and Capterra took 7.2 percent.

One number inside the study is worth looking at on its own. Ten Speed says that comparison-format queries produced 26.7 percent of all citations from only 20 percent of the prompts, which is the largest ratio of any single category and points to a type of page most B2B sites still spend too little on.

Why the split may look different for you

Search Engine Journal questioned Ten Speed on six points before publishing, including a prompt-count mismatch, how the comparison category was labelled, the sample size across distinct brands and verticals, whether the 24 percent versus 4 percent gap was statistically tested, and whether any of the citations connected to actual conversions. The agency answered each, but the underlying caveat remains that this is not a random sample of the B2B web.

The finding also applies specifically to the evaluation stage, when a buyer already knows the category and is comparing options. Awareness-stage prompts, where the buyer does not yet know a product exists, still rely much more on third-party content, so a team that reads this study as permission to stop investing in Reddit or YouTube would be reading it too narrowly.

The practical takeaway is that a B2B product page, comparison page, and how-to guide all matter more when the buyer decides than the AI-search discussion of the past year has suggested. A team weighing where to spend the next content week has one specific number to point at: comparison and how-to formats earned more citations than their share of prompts would suggest for Ten Speed's clients, and both are pages the product team controls fully.

Source

Product Pages Get 24% Of AI Citations, Reddit & YouTube 4%, Search Engine Journal by Greg Jarboe, 3 September 2026
What AI cites at the B2B evaluation stage, Ten Speed by Nelson Brassell, 3 September 2026

This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.