Skip to content
AI-Assisted Content6 min read

72% of AI Answers Skip Brand Names, Research Shows Which Content Type Actually Earns Citations

Publishing more content does not earn more AI citations. A new large-scale study published this month makes that uncomfortable truth concrete — and hands content teams a specific map of what actually works.

Conductor's AI Brand Recommendation Study, covered this week by the Content Marketing Institute, ran 14,000 API calls across 10 industries, seven intent types, four large language models, and five simulated user personas. The finding that should stop every content calendar review in its tracks: between 45% and 72% of educational queries — the intent category most content teams invest in heavily — returned zero brand names. Not fewer brands. Zero.

The predictor of AI brand visibility is not content quality. It is not publishing volume. It is query intent.

Intent Type Ranked From Most to Least Consistent

Conductor's study ranked seven intent types from least to most consistent in AI brand recommendations: Purchase, Support, Navigation, Recommendations, Pricing, Education, and Comparison. The ranking held across all four engines tested — ChatGPT, Perplexity, Claude, and Gemini — and across all 10 industries.

That stability is the finding. It doesn't matter whether a brand competes in insurance or cloud computing. The intent type of the query shapes how consistently AI recommends any brand within it.

Comparison sits at the top. AI picks a winner in comparison prompts and holds that position 91% of the time. The same brand leads the response across repeated runs. For brands that hold the top Comparison slot, that position is nearly unshakeable. For brands that don't, it is the hardest intent type to break into — precisely because the engine has already formed a stable opinion.

Purchase sits at the bottom. When a buyer is ready to make a decision, AI recommendations become a coin flip. Only 40% of brands overlap between any two runs of the same purchase-intent prompt. The consideration set reshuffles almost every time.

The Educational Content Paradox

Educational content has the second-highest brand overlap score in the study — meaning AI is reasonably consistent about which brands it considers authoritative in educational contexts. But that agreement almost never surfaces as a brand citation.

In 45% to 72% of Education prompts, depending on the LLM, the engine responded without naming a single brand. High agreement. Near-zero mentions.

"The goal of educational content isn't citations. It's getting LLMs to view your domain as an authority in other intent areas," Conductor's research notes. Teams measuring educational content by citation volume are measuring the wrong output from it.

The mechanism is less direct: educational content trains the model to associate a domain with a category. That association then influences recommendations at higher-intent stages — Comparison, Recommendations, Purchase — where brand citations actually happen. Skip the educational foundation and the downstream recommendations become unstable.

Internal links from educational content to pricing and comparison pages make that authority transfer explicit. AI crawlers follow those pathways when constructing recommendations for higher-intent queries.

Pricing Transparency Has a Measurable Gap

One in five pricing queries returned no brands at all across the Conductor study. Not because AI chose not to recommend anyone. Because the pricing data was not accessible enough to parse.

Machine-readable pricing tables close that gap. Image-based tables do not feed AI answer engines. HTML tables do. Shannon Vize, content marketing director at Conductor, speaking with the Content Marketing Institute this week, put it plainly: "Any detail you can provide instead of not having any pricing available is better than nothing."

If a pricing page hides numbers behind a demo request form, AI cannot confirm the answer and defaults to brands whose data is openly structured. Starting costs, price ranges, standardized tier breakdowns — all of it counts. The format matters as much as the presence of the information.

Comparison Content Rewards Honesty About Competitors

The conventional rule in brand marketing — don't mention competitors by name — actively works against AI brand visibility in comparison queries. The brand leading 91% of comparison responses in Conductor's study earned that position by being cited in structured, data-rich comparison content, not by avoiding competitor mentions.

Vize's guidance from the CMI interview: create comparison pages with accurate feature-by-feature tables and verifiable differentiators for two to five competitors. Then acknowledge where competitors hold a genuine advantage. Honest comparisons signal the kind of authoritative sourcing that AI engines prefer for comparison queries. Purely promotional comparisons do not.

The same brand leading 91% of comparison runs is not coincidence. It is a content structure that AI has ingested, scored for credibility, and returned consistently because the underlying data supports a stable verdict.

The Content Stack That Builds Stable AI Presence

Conductor's research describes a sequence that compounds. Education content trains the model to associate a domain with a category — even when brand names don't appear in outputs. Recommendation prompts, where AI chooses freely without any brand named in the query, reward domains that have built that foundational association. Comparison content is where stable brand positions are established. Pricing content closes the 20% gap of queries that return no brands at all because data was unavailable.

Each layer builds on the previous one. A site publishing only educational content accumulates domain authority but generates minimal AI citations. A site publishing only comparison or bottom-funnel content skips the foundation that makes those recommendations credible to the engine.

HubSpot's State of AEO 2026 report adds a layer to this: AI platforms show distinct content type preferences. ChatGPT favors comparison content. Gemini and Google AI Overviews most often cite blog posts. Perplexity prefers product listings and landing pages. Building across intent types also builds across engine preferences — the brands with stable multi-engine presence are the ones with the broadest content mix.

What This Means for Content Teams Operating at Scale

The research reframes what a consistent publishing schedule is actually for. It is not feeding a content volume metric. It is progressively covering the intent stack — educational coverage establishing domain authority, comparison and pricing content establishing citation eligibility, recommendation-intent content building unprompted visibility.

For marketing teams using AI-assisted tools to maintain publishing cadence, the Conductor findings give a specific content typology to work from. Not "publish more" but "publish by intent layer." An industry news section covering educational topics builds the foundation. Comparison and pricing pages convert that foundation into stable AI brand mentions. The human layer — fact-checking, brand voice, and original analysis — is what lifts AI-generated drafts into content AI engines treat as citable rather than derivative.

The 72% no-brand figure is not a reason to stop publishing educational content. It is a reason to map every piece of content to an intent stage, understand what each stage is actually supposed to produce, and measure accordingly. The brands earning stable AI recommendation positions are the ones treating content strategy as a compounding system rather than a traffic tactic.

Sources: