Barely. Traditional Domain Authority — the third-party, backlink-based score from tools like Moz or Ahrefs — is a weak predictor of whether AI engines cite you. In Featured's 2026 analysis of 22,881 AI citations across ChatGPT, Claude, Perplexity, and Gemini, 34.5% came from sites with a Domain Authority under 40. AI engines retrieve and cite at the passage level, not the domain level, so a small site with the cleanest, best-matched answer can be lifted ahead of a household-name domain.
Last reviewed: 26 August 2026. The headline figure is from Featured's AI Citation Report (405 GEO audits conducted 2 June–21 August 2026). It is single-vendor, preview-period data — attribute "Featured, 2026" and treat the exact percentage as directional. The corroborating correlations below are dated and attributed in place.
Does a high Domain Authority score get you cited by AI?
Not reliably. Domain Authority (DA) is a third-party score — most familiar as Moz's 0–100 metric, with Ahrefs' Domain Rating a close cousin — that estimates how likely a domain is to rank in classic Google search, built mostly from the quantity and quality of its backlinks. It was never a measure of AI citation, and the 2026 data shows it does not stand in for one.
Featured's AI Citation Report is the freshest cut. Across the audits it carried out during its preview period, it recorded the Domain Authority of every cited source:
| Featured AI Citation Report (2026) | Figure |
|---|---|
| AI citations analysed | 22,881 |
| GEO audits (2 Jun–21 Aug 2026) | 405 |
| Engines covered | ChatGPT, Claude, Perplexity, Gemini |
| Citations from sites with DA under 40 | 34.5% |
Source: Featured, 2026 (single-vendor; directional). More than a third of the cited sources were domains a classic SEO audit would treat as low-authority. That is the opposite of what you would expect if AI engines simply cited the biggest domains — and it lines up with the wider pattern that AI citations do not track Google rank either.
Why doesn't Domain Authority predict AI citations?
Because AI retrieval operates on different inputs from the ones Domain Authority summarises. DA is a proxy for link-based ranking power. An answer engine, by contrast, fans one question into many sub-queries and then selects passages that score on relevance to the sub-query, extractable structure, freshness, and corroboration — not on how many sites link to the domain.
The practical consequence: the unit that gets cited is a self-contained chunk, not a whole domain. A niche site that answers one specific question cleanly, in a table or an answer-first paragraph, presents a better candidate passage than a high-DA page that buries the same fact three scrolls down. This is why one 2026 correlation analysis reported Domain Authority explaining only about 3% of the variation in AI citations (single-source, directional) — the metric is measuring the wrong thing for this job.
What actually drives AI citations, if not Domain Authority?
The signals that do correlate are about your standing as a recognised source and the quality of the passage, not your backlink count. Here is how the better-evidenced levers compare, each attributed:
| Lever | What it is | Evidence (attributed, dated) |
|---|---|---|
| Entity authority | How strongly AI associates your brand with a topic | Repeatedly found to predict citations more than backlinks (2026) |
| Brand mentions across the web | Being named, with or without a link | Ahrefs: correlated with AI Overview presence 0.664 vs 0.218 for backlinks (75,000 brands, 2026) |
| Topical depth | Comprehensive, single-topic coverage | Focused explainers overtake broad guides on citations (Promptwatch, 2026; directional) |
| Extractable structure | Answer-first passages, tables, clean HTML | Zyppy rated "answer near the top" 8.8/10, "AI-ready structure" 8.6/10 (2026) |
| Freshness | Recently reviewed/updated | AI-cited pages average ~25.7% fresher than organic top-10 (2026) |
Source detail and the ranking of these levers are laid out in our three AI-citation levers reference. Read every figure as directional and correlational — but note the through-line: none of them is Domain Authority, and the one that looks closest ("authority") is entity authority, a different thing.
Domain Authority tells you how a page might rank in Google. It does not tell you whether an AI will cite it — a third of AI citations come from sites a classic audit would call low-authority.
Does this contradict "authority matters for AI citations"?
No — the two claims are about different kinds of authority, and conflating them is the whole trap. When the research says authority is a strong AI-citation signal, it means entity authority: how well-established and credible your brand is as a recognised entity for a topic, read from brand search volume, knowledge-graph presence, and how often your brand co-occurs with its category. That is earned by being talked about, not by accumulating links.
Traditional Domain Authority is the backlink-based SEO score. The same Ahrefs data that props up "authority matters" is the data that separates the two: branded web mentions correlated with AI presence about three times more strongly than raw backlink count. So the honest reconciliation is that being a recognised source matters a great deal, and your Moz or Ahrefs authority number barely moves the needle on its own. High entity authority usually comes with a high DA, which is why the two get muddled — but it is the mentions and the topic association doing the work, not the link score.
What should you do about your Domain Authority for AI visibility?
Stop treating the DA number as your AI-visibility scoreboard, and spend the effort where the citations actually come from.
- Do not chase the DA score for its own sake. Link-building to lift a third-party authority number is one of the slowest and least direct ways to earn an AI citation. It is not useless for classic SEO — it is just the wrong target here.
- Build entity authority instead. Earn brand mentions across independent coverage, keep your entity consistent across the web, and become the recognised source for a specific topic — the durable lever in entity strength.
- Win at the passage level. Answer-first chunks, facts in tables, clean server-rendered HTML — make each page a better candidate passage than a bigger domain's. Start with how to get cited by AI and prioritising your structured data.
- Check for the cited-but-not-named gap. Featured also flagged cases where an engine cites a brand's own site without naming the brand in the answer — the brand-mention vs source gap worth auditing.
- Measure citations, not authority. Track your own citation share per engine over time; a domain-authority tool cannot tell you whether ChatGPT or Perplexity actually cited you.
The honest read for late 2026: a low or mid Domain Authority is not the barrier to AI citation it is for classic ranking — a third of AI citations come from sites under DA 40 — so the winning move is to be the clearest, best-corroborated source on your topic, not the biggest domain. Watching which of your pages actually get cited, per engine and over time, is exactly what Buffy Intel measures. Questions: [email protected].