Backlinks Aren’t Dead — But 0.664 vs 0.218 Explains Why AI Search Picks Brands, Not Link Profiles

Branded web mentions correlate far more strongly with AI visibility than backlinks do: 0.664 versus 0.218, according to Ahrefs. That’s the cleanest explanation yet for why AI visibility is shifting from pure link authority toward discoverability, selectability, and entity trust.

Picture the old SEO playbook: build links, climb rankings, collect clicks. That still helps, but AI search has inserted a second decision layer between ranking and visibility.

• Backlinks still matter because they help pages get crawled, indexed, and retrieved
• Mentions matter more now because AI systems need brands they can recognize across the web
• Clear passages matter too because models quote extractable text, not just high-ranking URLs
• AI citations are selective in ways classic rankings never were

The shortcut broke. And that changes what off-page strategy is actually for.

The Death of the Backlink Proxy

And once you see that shortcut break, the old authority proxy starts to look shaky. The biggest surprise in this shift is not that backlinks lost value, but that they became the weaker signal.

According to Ahrefs’ AI brand visibility correlation study, branded web mentions show a strong correlation with AI visibility (0.664), while backlinks show a much weaker correlation (0.218).

• Classic SEO trained teams to use links as shorthand for authority
• AI engines appear to prefer brands that are mentioned repeatedly across the web
• A link graph alone says less about whether a model will trust a brand enough to cite it
• Mention strength signals identity in a way raw backlink counts often do not

In AI search, links can get you considered. Mentions help get you chosen.

That distinction sets up the more important question: if links are not the finish line, what exactly are they doing now?

The RAG Pipeline: Why Ranking #1 Isn't Enough

The answer is architectural. AI search is not separate from traditional search; it sits on top of it, then applies a second filter.

I’d frame it this way: SEO still handles the plumbing. It helps your page enter the candidate set, but that does not guarantee the model will surface your brand in the final answer.

• Layer one is retrieval through search indexes, crawlers, and live web access
• Layer two is selection where the model evaluates passages it can summarize or quote
• A page can rank well and still fail if its key information is buried or vague
• A lower-ranked page can win if it answers cleanly, clearly, and in the right context
• The model cites passages because passages are what it can actually use

That is the nuance most “SEO is dead” takes miss. SEO still matters, but mostly as the system that gets you into the room.

The Two Gates of AI Visibility

And once you think in layers, AI visibility becomes a two-gate system. First the engine has to find the page. Then the model has to feel confident selecting it.

That second gate is where many teams lose. Research from Moz’s AI Mode citation study found that nearly 9 in 10 Google AI Mode citations (88%) sit outside the organic top 10.

• Discoverability answers one question: can the engine retrieve your page at all?
• Selectability answers the harder one: is the page quote-worthy when retrieved?
• Organic rank and AI citation no longer move in lockstep
• Passage clarity becomes competitive because AI systems extract snippets, not full site reputations
• Fresh, specific answers travel farther than generic “ultimate guide” copy

A page can be visible to search and still invisible to the answer.

So if the real bottleneck is selection, off-page strategy has to evolve with it.

The Entity Authority Checklist

That evolution looks less like link building at scale and more like entity engineering. If AI systems cite brands they can identify, then your off-page work needs to make your brand easier to recognize everywhere.

Semrush’s backlinks and AI search study supports the middle ground here: backlinks still influence AI-generated mentions, but the gains are stronger when those links come from higher-authority domains, not just higher volume.

• Use one consistent brand name across your site, profiles, and third-party mentions
• Build repeated topical association so your brand keeps appearing near the problems you solve
• Prioritize authoritative mentions on trusted sites, podcasts, communities, and reviews
• Publish quote-ready pages with definitions, frameworks, stats, and concise explanations
• Tighten schema and FAQ readiness so machines can parse your content more cleanly

For Scout7 users, this is exactly where AI visibility becomes practical: not “more backlinks somehow,” but a cleaner, stronger brand entity the model can repeatedly validate.

Measuring AI Visibility Before Your Traffic Report Does

Once off-page strategy changes, measurement has to change too. Waiting for referral traffic alone means waiting until demand has already shifted.

That matters because discovery behavior is already moving upstream. According to Adobe’s 2026 AI and Digital Trends report, 1 in 4 customers globally (25%) now use AI platforms like ChatGPT as their top research tool.

• Citation share is an early signal of whether your brand enters AI-led buying journeys
• Traffic reports lag behavior because many AI impressions happen before the click
• Brand recall can rise first even when analytics platforms show little immediate referral lift
• Prompt monitoring matters because different questions trigger different citation winners
• AI visibility belongs upstream with demand creation, not only downstream with sessions

If buyers start research in answer engines, the smart move is obvious: track whether your brand appears before competitors do.

The New Off-Page Playbook

That leads to the practical reset. Don’t abandon SEO. Stop treating backlinks as the finish line.

The scale of this shift is already measurable. Semrush’s AI Visibility Index is built from more than 126 million real US AI search prompts across 22 industries and four AI platforms, which tells you AI-native measurement is no longer experimental.

• Keep the plumbing healthy with crawlability, indexation, rendering, and technical SEO
• Strengthen entity consistency across website copy, profiles, PR, reviews, and communities
• Design pages to be quoted with short definitions, strong subheads, and extractable passages
• Earn better mentions, not just more links from sources models are likely to trust
• Track AI citations directly so your team can see selection, not just ranking

For small teams, this is the more durable strategy. Build discoverability to enter the pool, then build selectability so the model picks you from it.

Frequently asked questions

Do backlinks still matter for AI visibility?

Yes. The article explains that backlinks still help pages get crawled, indexed, and retrieved, which means they still support discoverability. The shift is that mentions and extractable passages appear to matter more when AI systems decide what to cite.

Why can a page rank well but still not get cited by AI search?

Because ranking is only the first gate. AI systems retrieve pages through search infrastructure, then apply a second selection layer that favors clear, quote-ready passages the model can summarize or cite.

What does 0.664 vs 0.218 actually mean?

In the Ahrefs study cited here, branded web mentions had a much stronger correlation with AI visibility than backlinks did. The article uses that gap to argue that AI search is more likely to surface recognizable, repeatedly mentioned brands than pages that rely on link signals alone.

What should teams measure if traffic from AI search is still low?

The article recommends looking upstream at citation share, prompt monitoring, and brand appearance in AI-generated answers. Referral traffic can lag because many AI impressions happen before a user clicks, so visibility can improve before analytics shows it.

References

• Ahrefs, “AI Brand Visibility Correlations” — https://ahrefs.com/blog/ai-brand-visibility-correlations/
• Moz, “AI Mode Citation Study” — https://moz.com/blog/ai-mode-citation-study
• Semrush, “Backlinks and AI Search Study” — https://www.semrush.com/blog/backlinks-ai-search-study/
• Adobe, “2026 AI and Digital Trends Report: Four Key Takeaways” — https://business.adobe.com/blog/2026-adobe-ai-digital-trends-report-four-key-takeaways
• Semrush, “AI Visibility Index” — https://ai-visibility-index.semrush.com/