Why are use-case pages easier for AI to cite?
Because the question a buyer asks AI is often a use case: "a tool for teachers that turns notes into lesson slides". When a page's title, first sentence and steps match that use case, AI can lift a whole passage instead of piecing it together from a generic homepage.
This is inference: we found no public statistic saying use-case pages are cited more than other pages. The only basis is how product pages are normally written: a homepage has to speak to everyone, while a use-case page speaks to one kind of person about one task, so it is easier to write as a sentence that answers the question directly.
What parts should a use-case page have?
The table below is our summary after taking apart real pages on the two sites below; it is not an official standard. The right column shows what they actually do.
| Part | What goes in | What the real pages do |
|---|---|---|
| Title and first sentence | One sentence: who, doing what, with which product | The title of Gamma AI Slide's teacher page is "How teachers can make lesson slides with AI from their own notes" |
| Steps | The settings and order specific to this use case | The teacher page has four steps, says to pick the "Lesson" scenario and the "Students" audience, and to check the outline in lesson-plan order |
| Content unique to this use case | Something no other use-case page has | MochiStamp's work page has 24 workplace phrases, each with a one-line drawing note |
| A real example, with its limits stated | Write only what was actually done; say "none" where it was not | The sales page says no sales proposal was ever recorded and uses a product-launch metrics table instead; the teacher page's example is a course-design overview, not a classroom lesson deck |
| Limits and reminders | What the AI or the product cannot do | The teacher page has "What the AI will not guarantee", the sales page has "What to check before you send it" |
| FAQ | 3 to 5, matching the structured data | MochiStamp's work page has three, including how it differs from the "Polite" set; the sales page has five |
| Links to neighbouring use cases | The use cases the reader will ask about next | MochiStamp's page ends with a dozen or so links to related phrase sets and seasonal pages |
How does MochiStamp's use-case page work?
mochistamp.com/en/phrases/oshigoto (opened 2026-11-06) groups sticker phrases by occasion: the same site has 10 sets under phrases (the URLs show kihon, teinei, oshigoto and others; we only opened and checked the Basic and Work pages), plus the seasonal pages under mochistamp.com/en/ideas. The Work page opens by stating "24 phrases for coworkers and work friends, following the shape of a workday", then lists sample stickers, a character note (a hamster with glasses and a little tie), the 24 phrases each with a drawing note, three tips and three questions.
It is one real page rather than ten copies: the 24 phrases belong to this set alone, the sample stickers were drawn for it, and the FAQ answers "how does this set differ from Polite". Whether the other sets (we only checked the title and description of the Basic set) use their own phrases rather than a swapped word is something we did not check page by page, so it is unverified. The page shows no update date, which is one thing it lacks.
How do Gamma AI Slide's use-case pages work?
gammaslide.com/ai-lesson-slides-for-teachers (page dated 2026-11-05) splits the one product into use cases such as teachers, sales, students, business reviews, conference talks, training and project proposals. The teacher page covers settings only a teacher would use: upload your own notes, pick the "Lesson" scenario and "Students" audience, check the outline in lesson-plan order, and use five prompt templates (one lesson, revision before a test, unit overview, parents' evening, lab or hands-on).
gammaslide.com/ai-presentation-for-sales is structured differently: the sales page gives a 10-slide proposal outline (cover, client situation, cost of the problem, proposal, how it works, proof, timeline, investment, risks, next steps) and a pre-send checklist. Both pages state the limits of their examples: the teacher page's example is a course-design overview, and the sales page says no sales proposal was recorded. The benefit is that AI will not attribute to you something you never did.
Why not generate pages in bulk from one template by swapping a word?
Because Google lists such pages as spam. In the spam policies on Google Search Central (viewed 2026-11-06, page undated), "doorway abuse" means pages made to rank for specific, similar queries, with examples including "substantially similar pages that are closer to search results than a clearly defined, browseable hierarchy"; "scaled content abuse" means generating many pages mainly to manipulate rankings, with examples including "using generative AI tools to generate many pages without adding value for users".
We have no official documentation to cite for how AI search products treat this, but a simple check (inference) works: swap the use-case word on the page for another one; if the whole page still reads fine, it has nothing unique and should not be published.
| Check | Pass | Fail |
|---|---|---|
| Word-swap test | Swap in another use case and the steps and example stop making sense | The whole page still reads fine after swapping a word |
| Unique content | A list, table or data that belongs only to this use case | Only the title and the opening sentence differ |
| Example | An example that was really done, with its limits stated | Invented customers, numbers or screenshots |
| Quantity | Each page has searchers and something to say before you make the next one | Hundreds of pages generated at once |
How do you build use-case pages step by step?
1. List your buyers' real use cases first
Look at the wording in search terms and customer emails and write it as "who + doing what". One use case, one page.
2. Find what is unique for each page first
It can be a list, template, data or test result for this use case. If you cannot find one, skip the page and merge the content into another.
3. Write it following the table above
Answer in the first sentence; steps, example, limits and FAQ all aim at this use case. For how to write it see How to write content AI will quote.
4. Keep FAQ structured data identical to the body
The questions and answers in FAQPage must be text that is visible on the page; see How to write structured data.
5. Link use-case pages to each other and to comparison pages
After reading a use case, the reader's next step is often comparing products; see How to write an "X vs Y" comparison page.
6. Date every page and look again after a few weeks
Ask AI the same set of questions, see which page is cited and which is not, and merge or add unique content to those that are not.
FAQ
How is a use-case page different from a feature page?
A feature page says what the product can do; a use-case page says how one kind of person gets one specific thing done with it, with steps and an example unique to that use case.
How long should a use-case page be?
There is no fixed length. It is enough to cover the first-sentence answer, the steps, one example, the limits and 3 to 5 questions; length follows from the unique content.
Can I use AI to generate use-case pages in bulk?
You can use AI to write them, but every page needs unique content. Google's spam policies list "using generative AI tools to generate many pages without adding value for users" as an example of scaled content abuse.
Can I invent an example I have not done?
No. Gamma AI Slide's sales page says outright that no sales proposal was recorded and shows only a related real example.
Should use-case pages go in the sitemap?
Yes. Put new pages in the sitemap and link to them from related pages so crawlers can both find them and see where they sit in the site.
Sources
- Google Search Central · Spam policies for Google web search （Viewed 2026-11-06; the page shows no update date）
- MochiStamp · Work LINE sticker phrases （Read 2026-11-06, page shows no update date）
- Gamma AI Slide · AI Lesson Slides for Teachers （Read 2026-11-06; the page is dated as updated 2026-11-05）
- Gamma AI Slide · How to make a sales proposal presentation with AI （Read 2026-11-06; the page is dated as updated 2026-11-05）
- kagentx evaluations: the data on the case study pages; each example gives its date.