There’s a pitch going around for business owners who want ChatGPT to recommend them. It goes like this. Someone owns a handful of cheap websites. Each site publishes “best [your category]” lists. You go first on every list, with a link back to your real site. Some versions add hidden code meant only for machines. Before long, the pitch says, ChatGPT and Google start naming you.
I looked into it, because if you run a company of 5 to 50 people, someone is going to offer you this if they haven’t already. Here’s what the public evidence says, and what I’d do instead.
Do AI answers really lean on “best of” lists?
Yes. That part of the pitch is true, and it’s why the idea sounds plausible.
Ahrefs studied ChatGPT’s answers to 750 top-of-funnel prompts across software, products and agencies. Sorted by page type, “best X” blog lists made up 43.8% of the links ChatGPT cited. Ahrefs also found that 35% of the cited lists sat on low-authority domains.
Keep two things in mind. Ahrefs sells AI-visibility tools, so it has a stake in this topic. And that 35% figure is, in my reading, why the trick looks like it should work. If small, unknown sites get cited, why not a few you control?
What happened when someone tested self-made lists?
They got cited, but in flickers.
Ahrefs ran the test on its own products. It published 34 self-promotional pages and tracked 9,886 AI answers across ChatGPT, Gemini, Perplexity and Copilot, from February 7 to May 31, 2026.
Mentions did show up. The citations were unstable, though. In about a quarter of cases, a page was cited once for a query on one assistant, then never appeared again.
Being cited also didn’t mean being recommended. When an answer cited one of their conference pages, 43% of those answers never mentioned the conference at all. They named competitors pulled from the same page.
The number I keep coming back to: for their established product, 94% of the new mentions came from third-party content. Pages other people wrote.
So yes, you can get mentioned with pages you made. Ahrefs measured where new mentions came from, not how long they lasted. In my reading, the ones worth counting on are the ones other people write.
What do Google’s own rules say?
Google’s spam policy describes this setup directly.
Google’s spam policies define link spam as creating links to or from a site “primarily for the purpose of manipulating search rankings.” One named example is excessive link exchanges: “Link to me and I’ll link to you.” A group of sites that exists to list you first and link to you fits that description, in my reading.
The same page covers scaled content abuse. That’s producing lots of pages mainly to manipulate rankings instead of helping people, and Google says that applies no matter how the pages get made, including with AI tools.
What happens to a site that breaks these rules? In Google’s words, it “may rank lower in results or not appear in results at all.”
I want to be careful here. Google’s page talks about the site that violates the policy. It doesn’t spell out what happens to the business being promoted. My reading, and it’s only my reading, is that a network built to link to you ties your real site to it. If that network gets flagged, I wouldn’t want my main domain sitting at the end of all those links.
Doesn’t everyone already rank themselves first?
A lot of companies do, and this is where the evidence pulls in two directions.
The same Ahrefs list study looked at 250 Google searches for “best X software” and found 169 of them (67.6%) featured a list where the publisher ranked itself first. Self-ranking is common, and Google hasn’t wiped it out.
On the other side, Lily Ray reported organic visibility drops of 29% to 49% on several large B2B and SaaS brands that run many self-promotional lists, starting in mid to late January 2026. One of her examples fell 49% between January 21 and February 2. Another was down 29% since mid-January.
Both data sets can be true at once. She frames the lists as a possible cause (“it seems possible”), not a proven one. Her own line on self-promotional lists is that “it works, until it doesn’t.”
Also notice who those sites are. They’re big companies with real brands, ranking themselves on their own blogs. A set of throwaway sites built to rank someone else is a different setup, and I haven’t seen public evidence that it holds up better.
What about the hidden schema add-on?
Skip it.
Some versions of the pitch add structured data, often called schema, that only machines can see. Google’s structured data rules are plain: “Don’t mark up content that is not visible to readers of the page.”
Even visible schema isn’t a shortcut into AI answers. Ahrefs tracked 1,885 pages that added schema and found AI citations barely moved: AI Overviews down 4.6%, AI Mode up 2.4%, ChatGPT up 2.2%.
The pitch next to the evidence
| What the pitch says | What the evidence says |
|---|---|
| AI answers pull from “best of” lists | True. By page type, they were 43.8% of the links ChatGPT cited in Ahrefs’ study |
| Your own lists will get you named | Ahrefs got mentions, but in about a quarter of cases a page was cited once for a query on one assistant, then never appeared again |
| Being cited means being recommended | 43% of answers citing an Ahrefs conference page never mentioned the conference |
| It’s safe for your real site | Google calls links made to manipulate rankings link spam. The risk to your own site is my reading, not a Google statement |
| Hidden schema helps | Google says not to mark up content readers can’t see, and added schema barely moved AI citations |
So what’s the honest way in?
Get onto other people’s real lists. That’s my reading of the two Ahrefs numbers above, the 43.8% and the 94%.
A made-up example: say you run a 20-person bookkeeping firm that works mostly with restaurants. The useful move is finding the handful of sites that already publish “best bookkeeping services for restaurants” and giving each one a real reason to look at you.
The rest of the list is plain work:
- Pitch the people who already publish “best of” lists in your space. Bring something useful, like a free trial, real data, or a quote they can use.
- Publish useful content of your own. If you write a “best” list, say you’re the publisher and link to real alternatives.
- Show up under your real name where your buyers already are, like LinkedIn, YouTube and podcasts. Answer questions. Don’t drop links.
- Clean up how you appear online. One bio, one set of links, a visible About page, and markup that matches what’s on that page.
- Measure it. Ask ChatGPT, Perplexity, Google AI Mode and Claude the same fixed questions every week and write down who gets named. You’ll see the flicker instead of guessing at it.
If I wanted to test this cleanly, here’s what I would run. Twelve weeks on a site I own, with my own content plus outreach for placements on other people’s lists, and the same set of questions tracked every week. No network and no hidden markup. At the end you’d know which mentions came from your pages and which came from other people.
What should you ask whoever pitches you?
- Who owns the sites, and who else is listed on them?
- Will they link to my real site?
- Who writes the lists, and who reviews them before they go up?
- What happens to my site if Google acts on the network?
- Can you show me mentions that come from people who aren’t you?
That last question matters most. For its established product, Ahrefs saw 94% of new mentions come from third-party content. My reading is that those are the mentions to build on, and a seller with none to show is selling you their own pages.
When a customer asks ChatGPT for the best option in your category today, who does it name?
If someone has pitched you something like this and you want a second opinion, I’m happy to talk it through.