Which of your pages earn AI citations? Start with your industry

Documentation made up 9.4% of owned citations in Information Technology and 0.9% in Sales and Marketing. Software had the largest comparison-page share among the displayed industries, at 23.4%. Sales and Marketing had the largest homepage share, at 15.7%.

Use these differences to choose what to inspect on your site. They describe the pages cited in each industry, not a publishing quota or a guaranteed route to more citations.

Across the full study, blog and editorial pages supplied 28.0% of owned citations and comparisons and alternatives supplied 18.6%. Product and commerce pages supplied 6.5%, and homepages 10.2%.

The overall mix is useful context, but a buyer’s task is the better starting point. Your product page, documentation and comparison page can each answer a different part of the decision.

In the measured owned-citation mix, docs and help accounted for 96.7% of Android’s citations, 74.7% of PyTorch’s and 74.1% of NVIDIA’s. Dedicated documentation properties also concentrated their citations in documentation pages.

A small global share can therefore hide an important job for a specific product. Make setup, compatibility and capability information clear where technical buyers expect to find it.

About a third of owned citations, 32.2%, pointed to addresses outside the nine named page groups. Many were short, custom paths rather than recognizable blog or product directories.

Those pages are included in the denominator. Removing them would make every named format look more dominant than it was. The useful next step is to read the pages that matter for your brand, rather than assume their role from the address alone.

Choose a few important buying questions and inspect the cited pages. Is the missing answer a capability explanation, a comparison, a setup detail or a cost calculation? Improve the page best suited to that job.

Use the industry benchmark to challenge an assumption about your content plan. Use the actual questions and sources to decide what to write.

The study groups addresses by subdomain and path, with the first matching rule assigning the page type. It combines both engines and does not provide an engine split. Industry is the brand’s top market. Platform sites whose owned citations are mainly user content and Parse’s own site are excluded.

The classifier labels 67.8% of links; all unclassified links remain in the overall denominator. These are shares of citations to brand-owned sites, not shares of all AI citations or estimates of format performance.