The short version: Domain Authority (DA) was a useful proxy for Google rankability in the backlink-driven era. In 2026, it has become the wrong metric to optimize for. The number that matters now is Citation Authority, which measures how often a domain gets cited by AI engines (ChatGPT, Perplexity, Gemini, Claude, Google AI Overview) when users ask questions related to your category. DA and Citation Authority correlate weakly, sometimes inversely. The work to build them is different. The teams that figure this out first capture the AI search wave.

The Salesforce State of Marketing 2026 report shows 85% of marketers say AI is reshaping their SEO strategy and 88% have begun optimizing for AI-generated responses on ChatGPT and Google's AI Overview. High-performing marketers (those getting the highest returns on marketing spend) are 2.2 times more likely than underperformers to have optimized for AI search. Yet most SEO dashboards still report DA as the headline metric. The mismatch between what teams measure and what AI engines actually reward is the 2026 story.

Why Domain Authority Worked and Why It Stopped

Moz launched Domain Authority in 2013 as a 1-to-100 logarithmic score based on backlink profile quality: referring domains, link equity flow, spam signals. It worked because Google's ranking algorithm leaned heavily on PageRank-derived signals. Sites with strong backlink profiles ranked, sites without them did not. DA was a proxy. Imperfect, but directionally useful.

LLM-based search does not work this way. When ChatGPT, Perplexity, or Google's AI Overview generates an answer, the system pulls from a different signal set: semantic match between the query and the source content, structural extractability (whether the answer is in a format the model can pull cleanly), factual specificity, source recency, explicit author credibility markers, and citation patterns in the model's training data and retrieval-augmented index. None of these correlate cleanly with DA. A page with DA 30 might earn citations from ChatGPT for a specific query while a DA 80 page in the same category gets ignored entirely.

The independent test any team can run: pick a query in your category. Search it on Google and note the top organic results. Then ask the same query in ChatGPT, Perplexity, and Google AI Overview. Compare which sources the AI engines cite. The overlap is usually 30 to 50%, sometimes lower. The sources cited by AI but not ranking on Google tend to share four traits: structured content, specific data points, recent publication dates, and explicit authorship.

What is Citation Authority?

Citation Authority is the measurable rate at which a domain gets cited by AI engines for queries in its target category. Unlike DA, no single vendor owns the metric. It gets measured through tools in the AEO (answer engine optimization) category. Profound, Peec AI, AIclicks, Ahrefs Brand Radar, SE Visible, and Conductor all measure variations of it. The Conductor 2026 AEO/GEO Benchmarks Report, drawing on 13,770 enterprise domains across 10 industries, found that ChatGPT drives the majority of AI referral traffic. The category is real, the measurement infrastructure exists, and the signals are auditable.

Citation Authority is built on four signals that DA does not capture:

Structured extractability. Pages that present information in formats LLMs can extract cleanly (definitions in the first 100 words, tables, comparisons, numbered lists, FAQ blocks) get cited more often. A 2,000-word essay with no extractable blocks can have high DA and zero AI citations.

Factual specificity. AI engines cite pages with specific, verifiable claims (real numbers, named sources, dated examples) more often than pages with general statements. "Marketing leaders prioritize AI in 2026" reads as a paraphrasable summary. "75% of marketers have adopted AI per Salesforce State of Marketing 2026 across 4,450 respondents" reads as a citation candidate.

Source recency. AI engines, especially those with real-time retrieval, weight 2025 to 2026 sources higher than 2021 to 2023 sources for queries about current state. A page refreshed monthly outperforms a page with strong DA but a 2022 publish date.

Citation in third-party listicles and roundups. AI engines lean heavily on aggregator content (top 10 lists, comparison articles, expert roundups) when generating answers about tools and services. Getting cited in those third-party pieces matters more than building DA on your own domain. The mention is the asset.

How RZLT Works On This In Practice

RZLT measures Citation Authority alongside (not instead of) traditional rankings. The May 2026 site consolidation work is a concrete example: 116 redirects shipped to clean up duplicate-title and orphan-page issues across 350 indexed pages, recovering approximately 60,000 monthly impressions in Google Search Console. That GSC data is one half of the picture. AI citation tracking is the other half. The same content infrastructure work (clean structure, consolidated topical pages, refreshed publish dates, schema markup, extractable answer blocks) improves Citation Authority directly because LLMs reward the same signals the consolidation produced.

The practical work shift: spend less time chasing high-DA backlinks for their own sake and more time earning placements in the third-party listicles AI engines cite. Structure on-domain content for extractability (FAQ blocks, definition-first openings, comparison tables). Cite verifiable 2026 primary sources rather than aggregator stats. Publish refresh dates explicitly so models can see recency. Use schema markup that explicitly identifies the author, publish date, and content type.

The Takeaway

Domain Authority is not useless. It still correlates with Google rankability and remains worth watching for traditional SEO. But it is no longer the headline metric for teams building for AI search. Citation Authority, which is measurable, observable, and increasingly the leading indicator of brand discovery, has taken that slot. The teams that update their measurement stack first will know which content actually moves the needle in the LLM era. The teams that stay anchored to DA will keep optimizing for a Google-only world that no longer exists.

For the strategy layer behind AI citation work, see RZLT's complete guide to answer engine optimization. For the measurement tools that surface Citation Authority directly, see RZLT's Top 10 AEO Tools for Tracking AI Search Visibility in 2026. For the broader argument that agentic SEO is a workflow rather than a product category, see RZLT's POV on why agentic SEO isn't a product category you buy.