Brand Mentions vs Backlinks: Which Matters More for AI Search?
In Ahrefs' study of 75,000 brands, branded web mentions correlated with AI visibility at 0.66 to 0.71, far above Domain Rating at 0.27 to 0.33, and backlinks were very weak 1. Semrush found authority still mattered after a threshold 2. Both are correlations, so build mentions first and keep earning quality links.
- Ahrefs found branded web mentions correlated with AI visibility at 0.66 to 0.71. Backlinks showed very weak correlations.
- Semrush found Authority Score correlated with AI mentions at 0.65 (Pearson), though less with share of voice, and mainly past a threshold.
- All of these are correlations. Neither study shows that mentions or links cause citations.
- Practical split: earn mentions where your buyers look, and treat links as a quality bonus, not a volume target.
Which matters more for AI search, brand mentions or backlinks?
Mentions show the stronger link to AI visibility in the best public data, but nobody has proven cause. Ahrefs studied 75,000 brands and found branded web mentions correlated with AI mentions at 0.664 for ChatGPT, 0.709 for AI Mode and 0.656 for AI Overviews. Domain Rating scored 0.266, 0.285 and 0.326, and Ahrefs called link correlations very weak 1.
Semrush studied 1,000 domains and reached a more mixed view. Authority Score, a measure built partly on backlinks, correlated strongly with AI mentions (Pearson 0.65), but only moderately with share of voice (Pearson about 0.23) 2. Microsoft's Bing team still lists backlinks among the basics that remain essential for discoverability 3.
So the honest answer is both, in different roles. Links help engines find and trust a site. Mentions help engines connect your name to a topic.
What counts as a brand mention and what counts as a backlink?
A backlink is a clickable link from another site to yours. A brand mention is any place your name appears in text on the web, with or without a link. A citation by an AI tool is a link in its answer pointing to your page.
The three differ in an important way. A mention can be unlinked, such as a review that names your company. A backlink can say nothing about your topic, such as a directory link. Our reading (not from a study): AI tools read text, so a mention in a relevant sentence may carry more meaning for them than a bare link.
What does the evidence say, study by study?
The table lists what each source measured and found. All studies are correlational, and two come from companies that sell SEO tools.
| Source | Sample | What it found on mentions | What it found on links |
|---|---|---|---|
| Ahrefs, December 2025 1 | 75,000 brands with DR above 40, tracked in Brand Radar | Branded web mentions 0.664 (ChatGPT), 0.709 (AI Mode), 0.656 (AI Overviews). Branded anchors 0.511, 0.628, 0.527. | Domain Rating 0.266, 0.285, 0.326. Backlinks and URL rating described as very weak. |
| Semrush, October 2025 2 | 1,000 random domains in its toolkit | Not measured as a separate factor. | Authority Score vs AI mentions: 0.65 Pearson, 0.57 Spearman. Vs share of voice: about 0.23 and 0.36. |
| Microsoft Bing, October 2025 3 | Guidance, not a study | Not covered in the passage we cited. | Lists backlinks among basics that remain essential for discoverability. |
| Ahrefs self-promotion experiment, 2026 4 | 34 pages, 9,886 AI answers | One new conference brand: 82% of its 72 new mentions came from answers citing the author's pages. One established tool (Brand Radar): 6%. One company's experiment, not a general rate. | Not measured. |
How far can you trust these numbers?
Not very far, and the authors say so. Ahrefs states that correlation does not imply causation and that the study finds patterns, not confirmed ranking factors 1. Its sample only covers domains with a Domain Rating above 40 and a top keyword of at least 800 monthly searches 1. It says nothing about small, new sites.
Semrush uses its own Authority Score, so results may not transfer to other measures, and it notes its 1,000-domain sample may not represent all industries 2. The two studies also used different engines and metrics. That is why they do not agree neatly.
A fair summary: big, known brands tend to be both mentioned widely and linked widely. A study cannot easily separate the two. Treat the numbers as direction, not instructions.
Do backlinks still help in AI search?
Yes, in a supporting role. Semrush found link quality mattered more than raw link count, and that the effect showed mainly once a domain reached higher authority tiers 2. Nofollow and follow links showed nearly identical correlations with AI mentions 2.
That fits how search works. Many AI tools retrieve pages through a search step first, so a page that cannot be found or trusted by a search engine has a weaker start. Microsoft calls crawlability, metadata, internal linking and backlinks essential for discoverability in AI answers 3.
The takeaway is to stop chasing link volume. A few links from relevant, trusted sites beat hundreds from random directories, and the data does not reward the latter.
Why might brand mentions matter so much?
AI tools build answers from text, so repeated mentions of your name next to a topic make you a likelier name to recommend. Ahrefs found branded web mentions had among the strongest correlations of the web-based factors it tested, with YouTube mentions around 0.737 overall 1. The YouTube figure counts titles, transcripts and descriptions that name the brand 1.
Ahrefs also ran an experiment on itself. It published 34 self-promotional pages and tracked 9,886 AI answers from February to May 2026. For one new conference brand (72 new mentions), 82% came from answers citing the author's own pages; for one established tool (Brand Radar), 6% did. This was one company's experiment, not a general rate 4. The test was small and run by a company with an interest in the result 4.
It also found a catch. In 43% of answers that cited a conference page, the answer never named the conference and recommended competitors instead 4. Being cited is not the same as being recommended.
How should you split your effort between the two?
Put most of your outreach time into getting named in the right places, and keep a steady trickle of quality link work. Our suggested starting point, not a tested ratio, is mentions first and links second, but the split should follow your gaps.
Begin with a check: search your brand and your category in AI tools and see who is named. Then fix what is missing. Use digital PR if you have no press coverage, and see Reddit versus your own blog for forum tactics.
- Mentions: expert quotes in trade press, podcasts, reviews on third-party sites, honest forum answers, YouTube videos that name you.
- Links: relevant industry bodies, partner pages, original data worth citing, a clean internal link structure.
- Skip: bulk directory links, paid link packages, fake reviews. None of the studies support them, and they risk your reputation.
How do you measure whether it is working?
Track two things monthly: how often your brand is named in AI answers for 10 to 20 buyer prompts, and which sites those answers cite. Our query fan-out explorer shows the searches behind a prompt, which hints at where to seek mentions.
Record the sources that appear most for your category. Those are the sites where a mention is most likely to help. Re-check after you have earned new coverage, and expect weeks, not days.
- Month | Prompts tested | Times we are named | Times we are cited | New mentions earned | New links earned
- Nov | 15 | 2 | 1 | 3 | 1
- Dec | 15 | 4 | 2 | 5 | 0
Frequently asked questions
Do backlinks matter for ChatGPT?
Less than mentions in the best public study. Ahrefs found Domain Rating correlated at 0.266 with ChatGPT mentions and called link metrics very weak, while branded web mentions reached 0.664 1. It is correlation, not proof of cause.
Are unlinked brand mentions useful?
Should I stop building links?
Can I pay for brand mentions?
We would not rely on it. None of the studies we opened tested paid placements, and fake or undisclosed ones risk your reputation. Earned coverage, genuine reviews and expert quotes are safer and match what the data describes.
Does this apply to small businesses?
Unclear. Ahrefs only studied domains with a Domain Rating above 40 and a keyword of at least 800 monthly searches 1. Small or new sites were outside its sample, so test the idea on your own prompts.
Sources
- Top Brand Visibility Factors in ChatGPT, AI Mode, and AI Overviews (75k Brands Studied) · AhrefsOpened October 20, 2026. Supports: 75,000 brands, DR above 40; Spearman correlations for branded web mentions, anchors, DR, YouTube mentions; link metrics very weak; causation caveats (12 December 2025).
- Do Backlinks Still Matter in AI Search? Insights from 1,000 Domains · SemrushOpened October 20, 2026. Supports: 1,000 domains; Authority Score vs AI mentions (Pearson 0.65, Spearman 0.57) and share of voice (about 0.23 and 0.36); quality over volume; nofollow similar to follow (16 October 2025).
- Optimizing Your Content for Inclusion in AI Search Answers · Microsoft Advertising (Bing)Opened October 20, 2026. Supports: Crawlability, metadata, internal linking and backlinks remain essential for discoverability (8 October 2025).
- Self-Promotional Content Works, Until It Backfires (AI SEO Experiment) · AhrefsOpened October 20, 2026. Supports: 34 pages, 9,886 answers, 7 February to 31 May 2026; 82% vs 6% of new mentions from the author's pages; 43% of cited conference answers recommended other brands; small single-author experiment.