How to Target “Best X for Y” Queries in AI Search

What does “Y” usually mean?

Each type of Y needs different evidence. List the ones that matter to your buyers, using sales notes, search queries and support tickets, before writing anything.

How do AI systems use third-party consensus for “best” queries, and what do my own pages do?

For a "best" question, assistants search, often running several related searches, and draw on the pages that rank for them. Those are mostly third-party: roundups, comparison sites, review platforms, analyst lists and forum discussions. Research on cited sources shows how heavily that consensus weighs. A study of citations for branded questions found reviews and other social proof made up 57% of cited sources, against 12% for product and commercial pages. An observational study of about 200,000 AI responses found that appearing in frequently cited third-party lists was strongly associated with being mentioned. Ranking first in those lists added 16.5 percentage points to the mention rate in B2B software, with the benefit levelling off after a few placements.

Your own pages do three jobs:

• They confirm fit. Once you are on the shortlist, assistants look for facts that match Y: integrations, compliance, pricing, features for that use case.
• They supply the description. What an answer says about you often comes from your page, so its accuracy depends on how plainly the page states the facts.
• They win narrow questions. For a specific Y with little third-party coverage, a clear use-case page can be one of the few relevant sources.

A list of "best" products on your own site that ranks you first can be cited, since Ahrefs found self-ranked "best X" lists among the most prominent ChatGPT sources, but a citation is not a recommendation and buyers trust such lists less. Several SEO analysts reported Google visibility drops in early 2026 for sites relying on self-ranked "best" lists, though Google did not confirm a specific change. The effect of focused pages is covered in whether focusing a page on one use case improves AI recommendations.

How do you find and earn the third-party sources?

1. Run the questions for each Y value in the assistants your buyers use, several times each.
2. List every cited source, and note which ones name your competitors but not you.
3. Rank them by how often they are cited.
4. Check each for how it chooses products: editorial testing, reader submissions, review data or vendor applications.
5. Pursue inclusion on merit: offer a product trial to editors, submit to lists that accept submissions, build your profile on the review platforms cited, and give writers accurate facts.
6. Avoid paying for placement without disclosure. It carries disclosure risks, and it tends to produce the kind of page buyers and editors trust less.

How to decide which sources to prioritise is covered in why a page can rank in Google but still not appear in AI answers.

Where do comparison pages fit?

Between the two fronts. A fair comparison page on your own site, covering you and the two or three alternatives buyers weigh for a given Y, gives assistants a structured source for the follow-up questions. Those are "how does [you] compare with [competitor] for [Y]?" and "which is cheaper for [Y]?". Keep it honest: say where a competitor is the better fit, name the basis of comparison, and keep prices and features current. A comparison page that only praises you is less useful to buyers and more likely to be cited for a competitor's entry. One that helps a buyer choose, including choosing someone else, gives assistants accurate facts about you and earns readers' trust. How to structure these pages is covered in whether you should publish comparison and alternative pages for AI search visibility.

How do you optimize for AI buyer questions like “Which supplier is best?”

Supplier questions in manufacturing, distribution and business services are "best X for Y" questions with procurement criteria as Y. Buyers ask things like "best supplier of food-grade stainless tubing in the Midwest" or "which contract manufacturer is best for small-batch medical devices". Answers look for facts that match the criteria:

• Capabilities: processes, materials, tolerances, capacity.
• Certifications: quality, industry and safety certifications, with certificate numbers and the issuing bodies.
• Commercial terms: minimum order quantities, typical lead times, sample policies.
• Coverage: regions served, shipping, local stock.
• Industries served, with examples.

Publish these on a capabilities page in text, not only in a downloadable brochure. Keep your listings in industrial supplier directories and trade association member lists consistent with it. Procurement questions often cite those directories, and a certification or capacity figure that differs between them undermines the answer.

How do you measure progress?

Keep a fixed set of "best X for Y" questions, several for each important Y, and run them on a schedule. Track:

• Mention rate for each Y.
• Position in the answer when mentioned.
• Accuracy of how you are described for that Y.
• Sources cited, and which of your target sources now include you.

Expect the third-party work to show results over months, and page improvements to show sooner for narrow Y values. How to write the answer-first sections on your own pages is covered in how to write direct-answer-first content for AI search.

What might an illustrative case look like?

Consider an illustrative applicant tracking software company that rarely appeared in answers to "best applicant tracking system for staffing agencies". The cited sources were three roundups and a review platform's category page. It appeared in none of the roundups and had few reviews in the staffing category. Its site had a general features page and no staffing page. It built a staffing agency page covering bulk candidate import, client portals, pricing for agencies and two customer examples. It offered trials to the three roundup editors, one of whom added it, and asked every customer for an honest review in their own words. Five months later, it appeared in about 40% of the staffing answers, usually described using its new page.

What do people ask about “best X for Y” questions?

Should we target every possible Y?

No. Choose the Y values where you genuinely fit and where buyers ask. Weak fits produce weak pages and inaccurate answers.

Do listicles on our own blog help?

Honest guides help readers. Self-ranked "best" lists are often cited, but a citation is not a recommendation: in Ahrefs' test, an assistant citing a self-promoting list could still recommend a competitor from it, and some sites relying on such lists lost Google visibility in early 2026.

How long does third-party inclusion take?

Roundups update on their own schedules, often a few times a year. Plan for months, not weeks.