Why get into these articles?
Because when AI recommends products, the list mostly comes from the pages it searched and cited, and most of those are “best X tools”, “X alternatives” and “X vs Y” articles. Ahrefs analysed 15,000 prompts in 2025-08 and found only 12% of the URLs AI cited ranked in Google’s top 10, so ranking well on your own site is not enough.
For the two sites kagentx has evaluated, the cited pages the AI read look like this:
| Site | Evaluated | Cited pages read | Of which list similar products | Of which mention it |
|---|---|---|---|---|
| GammaSlide | 2026-10-26 | 28 | 19 | 0 |
| MochiStamp | 2026-10-22 | 10 | 10 | 0 |
| Total | 38 | 29 | 0 |
Across the two evaluations the AI read 38 cited pages; 29 listed similar products and 0 mentioned the evaluated site. “Listed similar products” is what this page means by list and comparison articles. This is two sites, one evaluation each, and does not speak for every industry.
GammaSlide
| Site of the page | Times cited | Products listed on the page | Mentions it | Contact options found on the page |
|---|---|---|---|---|
| creati.ai | 3 | 20 | No | Form, Contact page, Email |
| slidesai.io | 3 | 15 | No | |
| genai.works | 3 | 3 | No | Form |
| ppt.ai | 2 | 6 | No | |
| presentermedia.com | 2 | 2 | No | Contact page |
MochiStamp
| Site of the page | Times cited | Products listed on the page | Mentions it | Contact options found on the page |
|---|---|---|---|---|
| noripon.blog | 2 | 7 | No | Form, Contact page |
| 人気ランキング.com | 1 | 8 | No | Contact page |
| ai-rise.net | 1 | 8 | No | Contact page |
| note.com | 1 | 6 | No | Contact page, Form |
| unicorn-blog.jp | 1 | 5 | No | Contact page |
Each table above is the 5 most-cited pages. For GammaSlide these 5 were cited 13 times in total; 3 list 5 or more products and 5 have a contact option. For MochiStamp the 5 were cited 6 times; 5 list 5 or more products and 5 have a contact option. Pages mentioning the evaluated site: 0 on both sides.
A contact option is a form, contact page or email the evaluation found on the page. It only means the author can be reached, not that they accept submissions.
How do you do it?
1. Find which list pages the AI cites
Write down 5–10 questions buyers would ask (without your brand name, e.g. “best X tool”, “alternatives to X”) and ask each 3 times in ChatGPT, Perplexity and Claude with web search on. Record the source URLs under each answer and sort by site and times cited. Pages cited often that list a run of similar products are your targets.
The kagentx evaluation can do this step automatically; doing it by hand works too, and the full method is in AI visibility audit checklist.
2. Decide which ones you can get into
Open each target page and handle it by type:
- A personal blog or small publication with a contact form or email on the page: you can write and ask them to consider adding you.
- The page says “submit a tool”, “contribute” or “Suggest a tool”: submit following its rules.
- Product directories or aggregators: submit your product page directly.
- A “best X tools” list obviously written by a competitor, who ranks themselves first: unlikely to add you, so leave it for last.
- A page that has not been updated in a long time: check the date on it and leave very old ones alone for now.
3. Make your own pages worth citing first
When the author opens your site, one screen should show what you are, who it is for and what it costs. Put the homepage and pricing text directly in the HTML and use AI crawler check to confirm AI crawlers can get in. For the troubleshooting approach see Why doesn't ChatGPT recommend my product?.
4. Prepare materials for the author
Authors will not do your homework. Prepare a pack they can copy straight into an article:
- One-line positioning: “Product name is …, for …, …”. Facts only, no “the best”.
- 3–5 screenshots: the product interface and one real output, noting the version and date.
- A comparison table against the other products on the list: only things that can be checked, such as price, supported formats and limits, with sources for competitor information.
- Something only you have: real data, a tool, a feature others lack, so the author has a reason to write one more line.
- Your contact details and URL.
5. Do it yourself, no mass mailing
Writing or submitting is your own job. In each message say which article you read and how your product differs from the specific products already in it. Do not send one author several messages or mass-mail hundreds of sites with the same template. kagentx only helps you find targets and prepare materials and does not contact anyone for you.
6. Re-test with the same questions two to four weeks later
Compare how many times out of how many you appear, and whether you are on the cited lists, not any single answer. Lists need time to update and to be found again by AI, with no fixed timeline.
FAQ
Can I pay to be listed?
Some sites sell list placements and usually label them “sponsored”, which readers and AI can see. Whether to pay is your call; follow the site’s rules. Either way, put only checkable facts in your materials and let the author decide whether it is worth writing about.
What if the author does not reply?
That is common. You can follow up once after a while and then let it go. Meanwhile put effort into your own pages: an honest comparison page may itself get cited for the same questions.
Should I appear on several lists at once?
Being mentioned on pages from several independent sources helps, because AI combines multiple articles. But each should be something the author chose to write; do not spread the same text across many sites.
I am already on the list. What else?
Check the content is right: price, features and links up to date. If it is out of date, write to the author and attach fresh materials.
How do I know which list pages matter most?
Look at which pages are cited repeatedly in the AI answers’ sources and how many products each lists. The “times cited” and “products listed on the page” columns in the tables are worked out this way.
Sources
- Ahrefs · Only 12% of AI Cited URLs Rank in Google’s Top 10 for the Original Prompt （published 2025-08-11, 15,000 prompts）
- SparkToro × Gumshoe · AIs are highly inconsistent when recommending brands or products （2026-01, 12 prompts run 60–100 times on each platform）
- kagentx evaluations: the data on the case study pages; each example gives its date.