The page type AI engines cite is different in every industry — listicles win 61% of B2B-tech citations, homepages 55% of local-services citations, so there is no one AEO playbook

What the study measured

The primary source is DeltaV Digital's AI citation study, published July 13, 2026. Over a 90-day window (April 14–July 13, 2026) the researchers logged 21,075 responses from five AI engines — ChatGPT, Perplexity, Gemini, Google AI Overviews, and Google AI Mode — and analyzed 25,337 citations, taking the top-30 most-cited URLs and domains per brand across eight industries. Instead of asking “who gets cited,” it asked a sharper question: what kind of page does each industry's citations actually point to? The answer is not the same page type everywhere.

Every industry has a different citation fingerprint

Read the last row against the first four and the point lands: across the whole portfolio no single page type breaks 24%, yet inside a given industry one page type can dominate past 50%. A blended “publish more articles” instruction is close to the average and wrong for almost every specific client — articles are the plurality overall but a B2B-tech brand's citations are concentrated in listicles it usually doesn't even own, and a local-services brand's are concentrated in its own homepage.

Two more findings worth pricing on

• Comparison pages punch far above their weight. They were only 4.1% of citations by volume, but earned the highest citation rate per retrieval in the study — 1.87 citations each. A small number of well-built comparison pages did disproportionate work.
• Reddit is nearly unavoidable. It appeared among the top cited sources for 7 of the 8 brands studied — an off-domain surface no client controls, cited across almost every vertical.
• Owning your own citations is not guaranteed. Own-domain citation share ranged from 0% to 74.7% (a university at the top). One B2B firm earned 0% of the top-30 citations in its own category — every answer about its space pointed somewhere else.

Why this matters for AI visibility

• There is no portable AEO checklist. The deliverable that earns citations in one vertical (a listicle placement for B2B tech) is the wrong target in another (the homepage for local services). An agency reusing one content playbook across a book of clients is optimizing the average, which fits almost none of them.
• Some of the highest-value citation surfaces are off-domain. With Reddit in 7 of 8 fingerprints and listicles dominating B2B tech, a large share of what gets cited is a page the client will never publish — it lives on a third-party listicle or a forum thread. That's earned-media and digital-PR work, not on-site SEO.
• A page-1 ranking still doesn't tell you the page type. Knowing a client ranks says nothing about whether AI engines cite its homepage, an article, or a comparison page in its category — and that distribution is what decides where the next dollar of content work should go.

What an agency should do with this

• Fingerprint each client's category before you scope content. Pull the page types AI engines actually cite for that vertical, then aim the retainer at that type — homepage authority for a local-services client, listicle placements for a B2B-tech client — instead of a default blog-post cadence.
• Budget for off-domain citation targets. If listicles and Reddit are where the citations are, some of the work is getting the client mentioned on pages they don't own. Price that as digital PR, separate from on-site optimization.
• Baseline per engine and per page type, then re-measure. The fingerprint is the map; the mention rate across engines is the score. Establish both before you promise to move either.

The takeaway

DeltaV's durable finding isn't any one share — it's that “get cited by AI” resolves to a different page type in every industry, and often to a page the client doesn't own. That kills the idea of a single AEO playbook and makes the honest first deliverable a measurement: which page types get cited in this client's category, and where the client actually stands as a cited source, engine by engine.

You can run the free 60-second check on any business and see its mention rate across ChatGPT and Perplexity as separate numbers — the earned-citation surface, per engine (the $79 audit adds Gemini and Claude, 25 questions sampled three times each, because one run is a sample, not a score). If you run audits across a book of clients in different verticals, the $249 agency 5-pack runs the full repeated-sample audit across five businesses, white-labeled — the per-engine baseline you need before you decide which page type to optimize for each one.