Every few years, someone declares link building dead. It didn't die when Google rolled out Penguin. It didn't die when E-E-A-T became a thing. And it isn't dying now that AI search is eating the web.
What is happening is more interesting than death. The job link building does has expanded. The signals it creates now matter in more places than ever before. And the brands that understand this will show up in AI-generated answers while their competitors wonder where their traffic went.
Here's what the data actually shows.
AI Search Still Runs on Authority, And Links Still Build Authority
The hot take right now is that AI search works differently from traditional search, therefore SEO signals don't apply. That's a half-truth that leads people to the wrong conclusion.
Yes, AI search works differently. But it still needs to figure out who to trust. And the mechanism for doing that looks a lot like what Google has been doing for 25 years.
76% of AI citations come from pages already ranking in the top 10 organically. That's not a coincidence. AI systems are pulling from the same trust networks that traditional search does. Pages ranking first in Google see a 33% citation rate in AI Overviews, nearly double the odds of just being somewhere in the top 10.
You can't build AI visibility on a weak organic foundation. And you can't build a strong organic foundation without quality links.
The Mechanism Has Shifted — Links Are Now Trust Proxies, Not Just Ranking Signals
This is where it gets nuanced. Links aren't doing exactly what they used to do. They're doing something adjacent and arguably more important.
ChatGPT, Perplexity, and Google don't rely solely on backlinks to infer credibility. They look at how often your brand is mentioned in high-trust contexts, whether there's a link or not. The shift is toward co-citation and co-occurrence. But here's the thing: high-quality link building has always been about earning mentions in credible, contextually relevant publications. That's still the play. The mechanics downstream just reach further now.
Ahrefs analyzed 75,000 brands and found that brand mentions correlate more strongly with AI visibility than backlinks alone (0.664 vs. 0.218). That sounds like a case against links until you realize that brand mentions and links aren't opposites. A well-executed link building campaign earns both. A link in an authoritative editorial piece is also a brand mentioned in a trusted context. These things travel together.
Editorial and expert-source backlinks correlate strongly with AI search visibility. Directory submissions and PBN links fail in GEO. The cheap stuff was always cheap. Now it's also useless for AI visibility. That's not a new problem for link building, it's a new reason to do it right.
The Data Has Some Real Nuance Worth Acknowledging
It wouldn't be a credible piece without acknowledging the counterarguments, so here they are.
One study found that when it comes to securing AI mentions and citations, content depth and readability matter most, while traditional SEO metrics like traffic and backlinks have little impact. And the direct relationship between backlinks and AI citations is weaker than with organic rankings, with a correlation of 0.37 versus 0.41 for organic keywords.
That's real. Backlinks aren't the only lever, and they may not be the most direct lever for AI citation specifically. Content structure, freshness, and entity density all matter. AI platforms cite content that's 25.7% fresher than traditional search results, and ChatGPT shows a strong recency bias, with 76% of its most-cited pages updated in the last 30 days.
So if you're only building links and ignoring content quality, you'll underperform. That has always been true.
What "Working Differently" Actually Means in Practice
The practical implication isn't "stop building links." It's "build links the way they were always supposed to be built, and pair them with a content strategy that AI systems can actually use."
The most successful brands are merging PR and SEO into one unified motion that generates earned media, high-quality mentions, and semantic signals AI engines use to surface answers. A link from a respected industry publication is simultaneously a ranking signal, a trust proxy for LLMs, a brand mention in a high-authority context, and proof to Google that someone credible thinks you're worth citing.
Original visuals and research pieces often get linked to naturally, especially by bloggers, journalists, and content creators. These assets are likely to be referenced by AI tools when answering factual, data-based queries. This isn't new advice. It's the same content-driven link building that has always produced the best results.
The game hasn't changed as much as people think. It's just that the stakes for doing it poorly are higher, and the payoff for doing it well now extends into a channel that barely existed two years ago.
The Bottom Line
Link building isn't dead in AI search. The brands screaming that it is are either selling you something else or haven't looked at the data.
What's changed is the bar. Low-quality links never deserved to work, and now they genuinely don't. Authority built through editorial coverage, original research, and genuine topical relevance matters more than it ever has because it now buys you visibility in both traditional search and AI-generated answers.
The question isn't whether to invest in link building. The question is whether you're building links that actually deserve to move the needle.
