Yes. Backlinks for SEO still matter, but they now do a different job. Links reach AI answers through rankings because Google still uses them to rank pages and AI assistants take most of their sources from pages that already rank. AI tools give more weight to how often the web names your brand, with or without a link. Put your budget into links that earn you rankings and brand mentions at the same time.

Backlinks still matter in AI search, mainly because they help a page get into the small group of results an AI tool reads before it answers. Google's AI Overviews pull their sources from the same index that ranks the blue links. In July 2025, Ahrefs found that 76% of AI Overview citations came from pages ranking in Google's top 10.
An Ahrefs update covering 863,000 keywords found that only 38% of cited pages also ranked in the top 10. Ahrefs puts part of the drop down to better citation detection and part to AI Overviews moving to Gemini 3 in January 2026. Siddharth Bhalsod's review of the Ahrefs data puts standalone assistants' overlap with rankings even lower, at about 12% from the top 10 for ChatGPT and Gemini. The figure also applies to Copilot.
So how much a high ranking helps depends on which tool your buyers use. Before you decide how much ranking is worth to you, type your own commercial queries into Google and into the assistant your customers prefer, then note who gets cited.
AI tools search an index to read the top pages, then cite the ones they used. According to Ahrefs, AI systems supplement their training data through retrieval augmented generation (RAG) by "pulling from search engine indexes." Retrieval-augmented generation (RAG) means the model looks up live pages before it answers.
Links decide which pages rank and therefore which pages the model reads. Once your page is in that set, the model judges it on the text: whether it answers the question clearly and whether other sources back it up.
The citation is also a link, pointing back to you. It sends you traffic and tells the reader where the claim came from. If AI answers mention your brand but never cite your pages, you're probably being described from other people's content. Look at which URLs the answers cite, and check whether those pages talk about you accurately.
The links that affect AI answers are links from pages that other sites already link to, and links whose anchor text names your brand. Ahrefs looked at the top 50 websites across 76.7 million AI Overviews and found a 0.70 correlation with mentions on highly linked pages. A brand named on one well-linked industry page shows up in AI answers more often than a brand with dozens of links from pages nobody links to.
The same Ahrefs work separates link types. In its study of 75,000 brands, branded anchors correlated at 0.527 with AI Overview visibility, while the raw backlink count reached only 0.218. Volume on its own does little. A link adds the most when it names you on a page that already carries weight.
| Link or mention type | Effect on Google rankings | Signal for AI answers | Typical cost (source, year) |
|---|---|---|---|
| Paid guest post | Weak to moderate, and it breaks Google's spam policies when paid | Weak unless the host page is well linked | $77.80 average (Ahrefs, 2024) |
| Paid niche edit (link added to an existing page) | Moderate, and it breaks the same policies | Depends on how well linked the host page is | $361.44 average (Ahrefs, 2024) |
| Digital PR placement | Strong when editorial | Strong, because it names the brand on a linked page | $300 to $500 most common (BuzzStream via Blazehive, 2026) |
| Link with branded anchor text | Normal link value | 0.527 correlation (Ahrefs) | No separate price |
| Unlinked brand mention | None as a link | 0.664 correlation (Ahrefs) | Earned through PR and community work |
Paid guest post
Paid niche edit (link added to an existing page)
Digital PR placement
Link with branded anchor text
Unlinked brand mention
Large language models (LLMs) pick sources based on how widely the web discusses a brand and on whether the brand's pages rank. Both are correlations, and neither proves that one extra mention causes a citation. Still, the pattern holds across tools, which is why it's a reasonable guide for where to spend.
Mentions are the strongest signal anyone has measured. In Ahrefs' study, branded web mentions correlated at 0.664 with AI Overview visibility, roughly three times the figure for backlinks. Brands in the top quarter for mentions averaged 169 AI Overview mentions, against 14 for the next quarter down. A follow-up covering ChatGPT and AI Mode found the same pattern, as it reported mentions correlating at 0.709 for AI Mode and 0.656 for ChatGPT.
Branded search volume correlated at 0.392 in the Ahrefs data, and the rankings in the first section above decide which pages the model reads. When people search for your name and your pages rank for the topic, the model has two separate reasons to trust you. Look up your branded search volume trend, then count how many of your target queries list you on page one.
Editorial links from relevant sites still move Google rankings, and they matter most in competitive markets. Backlinko analyzed 11.8 million Google results and found that the #1 result has 3.8x more backlinks than positions 2 to 10 on average.
Google itself says links count for less than they once did. Gary Illyes, an analyst on Google's Search team, said on a podcast that you "can absolutely rank without links," then added: "if you are in a competitive environment the links start to become significantly important." He also noted that ranking without links is harder in languages where the web has less content.
Paid links are the exception that hurts. Buying links to influence rankings breaks Google's spam policies, yet an Authority Hacker survey cited by Ahrefs found that 74.3% of link builders pay for them anyway. If a link exists only because money changed hands, assume Google can ignore it or penalize you for it.
No fixed number applies to every site. The number you need is the number of referring domains the pages ranking for your query already have. Without any links, though, your odds of getting traffic are poor. Ahrefs data summarized by Ziton Digital shows that 96.55% of pages get zero Google traffic, and only 1 in 6,671 pages with no backlinks clears 1,000 monthly visits.
Work out your own target in this order:
A small gap means content fixes will probably beat link building. A gap of 50 domains on a head term is a project measured in quarters, not weeks.
The links AI assistants notice come from pages that name your brand and already earn links of their own. When you choose targets, check the site's authority score alongside how well linked the host page is. Check whether the page names you in the text as well.
These off-page SEO methods earn links and mentions at once:
Each method still counts when it produces a mention without a link. An unlinked trade article that names your product adds to the signal AI tools weigh most despite its limited value for rankings.
You measure backlink impact on two scoreboards now: Google rankings and AI visibility. A link building program that only reports referring domains misses half the result.
For rankings, track new referring domains to each target page alongside that page's position for its query. If the position hasn't moved after 3 months, the links were probably too weak or the content gap was bigger than the link gap. For AI visibility, run a fixed set of prompts in Google AI Overviews and in the assistants your buyers use, every month. Record whether your brand appears and which URL is cited, and compare that with the mentions you earned in the same period.
The two numbers won't always move together. That's expected, because Ahrefs' own data shows ranking and AI citation have drifted apart.
Start with the referring-domain gap on your 10 revenue queries, because it tells you whether links or content are holding you back. Then choose link targets on 2 tests: the host page already earns links, and it names your brand. That one rule covers backlinks for SEO and visibility in AI answers, so one budget can serve both. Re-test your prompts every month, because AI citation patterns changed noticeably between Ahrefs' 2025 and 2026 studies.
Snoika audits websites and tracks brand visibility in Google and in AI assistants, with human review on top of the platform. To have your link gap and AI visibility measured together, request a proposal.
Links from pages that already attract backlinks, especially those using your brand name as the link text, are associated with AI visibility. Ahrefs found a stronger correlation for branded link text than for total backlink count. That doesn't prove a link will cause a citation.
Brand mentions across the web and search rankings are associated with AI source selection. Ahrefs also found a correlation with searches for a brand's name. These findings can guide priorities, but they don't establish a formula that guarantees citations.
There's no fixed target for backlinks for seo. Compare the number of unique websites linking to your page with the median for the top five pages ranking for your query. Exclude paid or spammy sources, and check relevance before using that gap as a planning benchmark.
Check AI visibility monthly and review ranking progress after three months. Track each target page's rankings alongside new linking websites. Use the same AI prompts each month to compare mentions and cited URLs, since ranking gains and AI visibility don't always move together.
AI answers may be describing your brand using other websites rather than your own pages. Check the cited URLs to see where the description comes from and whether it's accurate. A brand mention doesn't necessarily mean the assistant used your website as a source.


