There is a playbook going around for getting your product into AI answers, and the core of it is sound. The engine builds its shortlist from third-party roundups, so get into the roundups. Find the lists the engine cites, find the editor, pitch them, get added. Every guide on the first page of search tells some version of it, and most of them are competent.
What none of them stops to do is read the byline.
When I run category questions through the engines and open the pages they cite, a large share of those “best of” lists turn out to be written by a company that is on the list. Not a publication. Not an analyst. A vendor, writing a roundup of its own category, and ranking itself first. The engine reads it, treats it as a source, and repeats the order. That changes what the playbook is worth, because you cannot pitch your way onto a competitor’s list, and the lists you can pitch are a smaller pool than the guides imply.
What the source list actually looks like
Take one category I checked in September, the AI security operations platforms. Ask the engine for the best options and look at what it was reading. Most of the pages behind that answer came from vendors in the category. One security company’s “best platforms” post put itself at number one. Another’s “top five” put itself at number one. The ordinary search results underneath told the same story, with a third vendor’s list of eight starting with its own product. Each of those posts is a reasonable piece of content. Each is also an advertisement with a table of contents, and the engine had no trouble using them.
It is not a security quirk. In marketing compliance software, a vendor had published its own “best brand compliance software” list, and the engine named that vendor in the answers to two separate questions. In consumer health testing, one of the pages the engine leaned on most was a “best test in the country” article published by a testing company, and that company came first in the answer. In business messaging, the answers were built mostly from agency blogs, a vendor or two, and a Reddit thread, while a well-funded, profitable company in the category appeared only as a cited source and never as a recommendation.
Look at your own category with this lens and the pattern is usually obvious inside ten minutes. It is the single most useful thing you can do before spending an hour on outreach, and it is a step the source audit should always include: for every list the engine cites, write down who wrote it.
The three kinds of list, and what each one is worth
Once you sort by author, the cited lists fall into three groups, and they call for completely different responses.
Vendor-written lists. A company in your category wrote it, and it almost always ranks its author first. You are not getting on it, and you should not try. No competitor is going to add you to the page whose entire purpose is to put them in front of you. Pitching these is where most listicle outreach quietly burns its time, because they look like the perfect target: they are cited, they are on topic, and your competitors are on them.
Commercial lists. Agencies, affiliate sites, and “software review” blogs whose business is the list itself. Some of these will add you if you pitch well. Some will add you if you pay. A few will add whoever has an affiliate program. These are reachable, and they work, but they are the most volatile sources in the answer. They get rewritten, re-ranked, and replaced, and the engine’s trust in them seems to rise and fall with how many other sources agree. Treat a placement here as a lease, not a purchase.
Editorial lists. Analyst coverage, trade publications, practitioner communities, and genuinely independent reviewers who test products and say what they found. These are the hardest to get onto, because you cannot buy them and the editor does not need you. They are also the ones the engine treats as the most credible, and the ones that tend to stay in the answer longest. In high-stakes categories they carry far more weight than everything else combined, which is the same mechanism behind why security answers lean on analysts rather than review sites.
The guides you will find treat all three as one target list ranked by authority or traffic. Ranked by what you can actually do about them, they are three different jobs.
The response to a vendor list is to write the honest one
If competitors’ own lists are feeding the answer, the move is not to fight them. It is to publish a better list of your own, and to make it the most honest one in the category.
That sounds like the same game, and it is not. A vendor list that puts its author first and describes every rival in a sentence of faint praise is easy for a reader to discount and, increasingly, for an engine too. A list that names the real options, says plainly what each is best at, states where your product is the wrong choice, and puts real prices and real limits in front of the reader is the page the engine has the least reason to discount. It is also the page a buyer is most likely to trust, which matters more than the citation.
This is the same principle as winning comparison queries by conceding. The vendor list says “we are the best at everything.” The honest list says “we are the best at this, and for that other job you want them.” Only one of those is a sentence an engine can safely repeat.
Where the outreach effort should actually go
Once vendor lists are off the table, the target pool shrinks, and that is good news. A shorter list means you can do it properly.
For commercial lists, pitch the editor with the one thing their list is missing. Not “please add us,” but a specific gap: the list is a year old and misses a category of buyer you serve, or it compares on a dimension you lead on. Offer the facts in the format their list already uses, so adding you is twenty minutes of their time, not two hours. Then check the page monthly, because these are the placements most likely to disappear without anyone telling you.
For editorial lists, stop thinking of it as a pitch at all. Analysts need briefings, trade publications need stories, practitioner communities need a person who is actually useful in the thread. Those take months and cannot be faked, which is exactly why they are worth more. The Reddit pattern is the clearest example: the engine trusts the thread because nobody in it is selling, and that only stays true if you do not start.
What I got wrong
When I first noticed this, I overclaimed it. Looking at one category, I saw a vendor that had written its own list and got named, and a second vendor that seemed to fit the same pattern, and I wrote it up as two examples. When I went back to check the second one properly, I could not verify that its list existed or that it fed the answer. One example was real. The other was a pattern I wanted to see.
That is the trap this whole method sets. Once you know vendor lists are common, you start seeing them everywhere, and “seeing” is not the same as opening the page and reading the byline. Every claim about who wrote a list, and whether that list fed a specific answer, needs the page open in front of you. The pattern is real. Each instance still has to be checked.
Where this leaves you
The listicle playbook is right that third-party lists decide the shortlist. It is wrong to treat them as one pool. Before any outreach, sort every cited list by who wrote it. Leave the vendor lists alone and answer them by publishing the most honest list in your category. Pitch the commercial lists with a specific gap and treat each placement as something you will have to maintain. Put your real patience into the editorial sources, because those are the ones the engine trusts and the ones that last.
You will send far fewer pitches. More of them will land, and the ones that land will still be there next quarter.
If you want the lists behind your category answer sorted this way, with a plan for which ones are worth the work, that is the engagement, and it runs on a documented method.
1 response