When AI search started driving meaningful traffic in 2024, a predictable question emerged among SEOs and content creators: do backlinks still matter?
The short answer is yes. The nuanced answer is that they matter differently — through a different mechanism, with a different set of priorities for which links to build.
I’ve been building links for authority sites for over two decades and testing AI citation patterns for the past two years. What I keep finding: a link from a relevant, authoritative source in your niche does double duty — it contributes to traditional rankings and strengthens entity recognition that feeds AI citation confidence. A link from a high-DA but topically unrelated site does less of both.
Here’s the honest breakdown.
How backlinks affected search — the traditional model
In Google’s traditional ranking algorithm, backlinks are the primary authority signal. The PageRank algorithm treats each link as a vote of confidence from one page to another. More links, especially from high-authority pages, accumulates more ranking power. The mechanism is direct: more quality links → higher rankings → more traffic.
This model is well-understood, heavily gamed, and still functional in traditional search. Sites with more high-quality backlinks generally rank higher for competitive terms.
How AI engines evaluate sources — and where backlinks fit
AI engines don’t use PageRank. When ChatGPT, Perplexity, or Google’s AI Overviews generate a cited response, they’re not consulting a backlink database. They’re doing something closer to an entity and content evaluation:
- Is this source a recognised, verifiable entity?
- Does this source have demonstrated topical authority on this subject?
- Is this content structured in a way that answers the query clearly?
- Is the author verifiable and credible?
- Does the content have schema markup that communicates its nature and authorship?
Backlinks don’t directly feed most of these evaluations. But they feed the first two — and those turn out to matter quite a bit.
The indirect mechanism: entity recognition
Google’s Knowledge Graph — the entity database that AI language models are partially trained on — is built in part from backlink signals. When established, authoritative sites link to you, Google’s crawler interprets those links as evidence that your entity is recognised by the established web. That recognition feeds into Knowledge Graph entry probability.
An entity in the Knowledge Graph has a verified record that AI engines can reference. An entity without one is evaluated primarily on content and schema signals alone — which is less reliable for the AI engine than being able to cross-reference against a known entity record.
The mechanism: high-authority backlinks → increased probability of Knowledge Graph entry → AI engines have a verified entity record for your source → higher citation confidence.
This means backlinks still contribute to AI citation authority — just not directly. They’re one layer removed, working through the entity recognition layer.
For more on the Knowledge Graph side of this, see how to get into Google’s Knowledge Graph.
The indirect mechanism: topical authority signals
The second mechanism is topical authority. When established sites in your niche link to your content, they’re signalling to search engines — and through training data, to AI engines — that your content is recognised as authoritative by peers in the topic area.
A link from a respected HVAC industry publication to your HVAC content is a peer-recognition signal for topical authority. It tells AI engines “this source is acknowledged by the established players in this topic” — which increases citation probability for queries in that topic.
This is why topic-relevant links carry more weight for AI citation authority than off-topic links from high-DA domains. A link from a general business directory with a DA of 80 contributes less topical authority signal than a link from an industry trade publication with a DA of 45 in your specific niche.
See topical authority for the full picture of how topical authority is built and evaluated.
What backlinks don’t do for AI search
To be clear about the limits:
Backlinks don’t directly cause AI citations. You can have a strong backlink profile and get zero citations in ChatGPT if your content isn’t structured for citation (no schema, no clear Q&A structure, no named author). And you can get cited in AI engines with a thin backlink profile if your content is well-structured, topically authoritative, and schema-complete.
Backlinks don’t compensate for missing schema. In traditional SEO, a strong backlink profile can overcome technical deficiencies — a site with 10,000 strong links can rank despite weak schema. In AI search, schema is a direct input into the citation evaluation. No amount of link authority compensates for missing FAQPage schema on question-based content.
Backlinks don’t override poor content. In traditional SEO, low-quality content can sometimes rank if it has enough link authority. In AI search, content quality and structure are primary citation criteria. Thin, generic content doesn’t get cited regardless of how many links point to it.
What this means for link-building strategy
If you’re building authority signals for AI search, the shift isn’t “stop building links.” It’s “build the right links, and build other signals too.”
Keep building topical, relevant links. Links from established sources in your specific topic area remain the most valuable form of link-building for both traditional rankings and AI citation authority.
Prioritise brand mentions alongside links. For AI citation purposes, an unlinked brand mention in a credible context contributes entity signals that are meaningful in their own right. See unlinked brand mentions for how to build these deliberately.
Don’t treat links as the only authority signal. Schema markup, entity recognition, topical content depth, and named author credentials all contribute directly to AI citation rates. A link-only authority strategy will produce sites that rank well in traditional search but underperform in AI citation.
Digital PR produces both links and mentions. Press coverage, podcast appearances, and expert commentary in authoritative publications generate both linked citations (valuable for traditional rankings) and unlinked mentions (valuable for entity signals). This makes digital PR especially high-value for AI search authority. See digital PR for AI citations.
The compound effect of links + schema + entity
The highest-citation-rate sites combine all three:
- Strong topical backlink profile → entity recognition and topical authority signals
- Complete schema markup → machine-readable authority declaration
- Verified entity (Knowledge Graph, consistent NAP, sameAs schema) → AI engine can cross-reference and verify the source
Sites that have all three consistently outperform sites with only one or two. A high-link-count site with no schema and poor entity recognition will lose AI citations to a lower-link-count site with complete schema and a verified entity — consistently.
This is the core argument of the authority signals for AI search overview: the signals compound, and link-building is one component of a larger system, not the whole thing.
The honest bottom line
Backlinks still matter. They matter for traditional rankings, which still drive significant traffic. They matter for entity recognition, which feeds AI citation confidence. And they matter as topical authority signals when they come from relevant, established sources in your specific topic area.
What’s changed is that backlinks alone aren’t sufficient for AI search authority the way they were sometimes sufficient (at least in terms of rankings) for traditional search authority. The content quality, schema, entity recognition, and topical depth signals that backlinks previously could occasionally compensate for now need to be in place independently.
Build links. Also build schema. Also build entity recognition. Also build topical authority through content depth. The sites that do all of these compound their authority in both search channels simultaneously.
Frequently asked questions
Do backlinks help with AI search citations?
Backlinks don’t directly cause AI citations the way they cause ranking improvements in traditional Google search. However, they matter indirectly: they contribute to entity recognition in Google’s Knowledge Graph, and they signal topical authority by demonstrating that established sources recognise your content. Link-building still matters for AI search — but the mechanism is different.
What matters more for AI citations — backlinks or content quality?
Content quality and structure matter more directly for AI citations. AI engines evaluate content based primarily on: does it answer the query well, is it structured clearly, is the author verifiable, and does it have proper schema markup. Backlinks contribute to the entity trust layer, but a well-structured piece from a low-link-count domain can outperform a mediocre piece from a high-DA domain in AI citation rates.
Which types of backlinks matter most for AI search authority?
Links from established, authoritative sites in your specific topic area carry the most weight — not because of raw link equity, but because they contribute most strongly to entity recognition and topical authority signals. A link from a relevant trade publication is worth more for AI citation authority than multiple links from general, high-DA sites outside your topic area.
Should I stop link-building if I’m focused on AI search optimization?
No. Link-building remains valuable — it contributes to entity recognition, topical authority, and traditional search rankings which still drive traffic. The change is that content quality, schema, entity recognition, and topical authority now matter alongside links. Stop treating links as the only authority signal — add the others, don’t subtract links.