Who can get your B2B SaaS into the best-of lists AI answers pull from?
Which lists to target, who runs them, and how to get in
| List type | Who publishes it | What earns inclusion | First move |
|---|---|---|---|
| Category roundups on company blogs | Usually a person at a company in your category, who maintains the post by hand | A clear reason you are the best choice for a stated purpose, backed by first-hand evidence the owner can check | Pull the pages the engines cite for your buyer prompts, rank them by how many answers cite each, and pitch the owners in that order |
| Review platforms (G2, Capterra) | The platform, from reviews by verified users | Reviews from actual users of your product. Both exclude employees and competitors, and Capterra also excludes anyone with a financial interest in the product | Ask a broad cross-section of real customers to review you, with any incentive offered regardless of the rating they give |
| Editorial trade roundups | Editors at the publication | No engine or platform documents this. Give the editor verifiable evidence for why you belong | Check that the engines actually retrieve the roundup for your prompts before you spend a pitch on it |
| Community threads (Reddit) | The people in that community | Authentic participation, with your connection to the product disclosed | Answer real questions in communities where you have a personal interest. Never post as a customer |
| Your own comparison page | You | A real reason for every pick, first-hand evidence, a visible author, and a plain sentence on who you are | Put the comparison table in visible text on the page, and say who you serve in one sentence |
Why do AI answers lean on someone else's list?
Because that is what they read. In our study of 192 AI answers to 8 agency-buying prompts on ChatGPT, Perplexity and Google AI Overviews (30 days to 20 September 2026), we classified 2,728 citations by page type. Ranked lists and roundups carried 1,570 of them (57.6%, across 342 distinct URLs). A vendor or agency's own site carried 852 (31.2%). Press release wires carried 1.0%, and trade press and news 0.1%.
A longer window says the same. Our 90-day category study counted 128,515 citations between 25 April and 24 July 2026, and listicles took 53.17% of them (68,329 of 128,515), more than every other page type combined.
Those shares do not mean lists are everything. Our benchmark of 160,240 citations across five B2B SaaS brands cut the data by domain type instead, and found that "company websites won half of everything AI cited (50.7%)". Different corpus, different classification. Both readings are true, and neither is a universal share of AI answers.
The harder gap is between being read and being named, or in the 90-day study's words, "You can be the library the model reads from and still not be the name it repeats." The name an engine gives first is shaped by how often a brand shows up across the whole set of pages it retrieved, including third-party lists the brand does not control.
We tested that on ourselves. In a week of tracked answers (17 to 23 September 2026, 2,895 answers on ChatGPT, Perplexity and Google AI Overviews), LoudFace was named 631 times, and 624 of those answers (98.9%) had read a loudface.co page. Of the 1,819 answers that read no LoudFace page, 7 named it (0.4%). Three of those were Perplexity answers that named us from other people's lists, so other people's pages can put you in an answer, just rarely.
So engines name you almost only when they have read a page that names you. Your own pages cover the prompts where the engine already retrieves you. For the generic category prompts, it usually retrieves someone else's list.
That is the case for third-party lists. They are the pages an engine reads when it has never heard of you. The same study calls them the slow route, to work once your own pages are in place.
Which sources does each engine read?
Different ones, and the differences are large enough to change your plan.
Start with depth. In the 192-answer study (64 answers per engine), ChatGPT read 21.5 distinct URLs per answer, Perplexity 14.2 and Google AI Overviews 6.9. A list has far more chances to land in a ChatGPT answer than in an AI Overview. Namings differed as sharply by engine in that study, and a blended figure hides which engine is the gap.
The source mix differs too. Our citation benchmark (30 days to 1 June 2026) classified sources by domain type, and found ChatGPT's user-generated share at 15.0% and Perplexity's editorial share at 13.6%, the highest of the three engines on each. Reddit was the most-cited single domain with 11,237 citations, and 80.5% of those came from ChatGPT (9,050 of 11,237). Treat AI search as one thing and, as the benchmark says, you are "averaging across machines that disagree."
Three list types stood out:
- LinkedIn articles, not company pages. In the 192-answer study, social citations came to 155. Of those, 149 came from 32 LinkedIn Pulse articles and 4 from ordinary posts. A long-form article on LinkedIn works as a list page in its own right.
- Reddit and Quora threads carried 41 citations (1.5%) in that study, and YouTube 15 (0.5%). In the 90-day study, Reddit was the second most-cited source overall with 5,903 citations.
- Review directories carried 13 citations (0.5%) across 4 URLs, all on ChatGPT. G2 never appeared as a source in any of the 192 answers. Those were agency prompts, so read it as one category's result. Software prompts may behave differently, and we have not measured them for review platforms.
Google says AI Overviews and AI Mode may be "issuing multiple related searches across subtopics and data sources", so one buyer question becomes many searches, and a list can come in through any of them. And a list can only be quoted if the engine is allowed to fetch it. OpenAI's crawler documentation says: "Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers, though can still appear as navigational links." Perplexity says PerplexityBot "is designed to surface and link websites in search results on Perplexity."
How do you find the lists that matter for your category?
Ask the engines, then look at what they read.
Write down the questions your buyers actually ask: "best [category] for [use case]", "[competitor] alternatives", "[category] tools for [company size]". Run each one on ChatGPT, Perplexity and Google AI Overviews, and record every page each answer retrieves. The pages that come back again and again are your target list. Our 192-answer study put the rule this way: "Pull the pages the engines actually cite for your buyer's prompts, count how many answers cite each one, and work the list in that order." In a corpus like that, the top 30 pages are "a finite, knowable target set".
Then look at who owns each page. In that study, six of the nine most-cited third-party pages were a list article on a company's own blog, one was a LinkedIn article and one a directory-style site. It was an agency-category study, and most were agencies ranking their competitors. "Most of them are somebody's blog post, and a person maintains it." That changes the job. You are rarely pitching a newsroom. You are writing to a named marketer at a company in your space, who updates the post when they have a reason to.
Before you pitch, check three things on each page:
- Open the live page yourself and search it for your name. A tracker can mark a page as read when all it got was a challenge page. That happens on sites that block crawlers, and in our 192-answer study it left presence on several cited pages unverified.
- Check the site's robots.txt. A list on a site that blocks OAI-SearchBot cannot be quoted in ChatGPT search answers, however good it is.
- Check which engine retrieves it. A page that only one engine reads is worth pitching only if that engine is where your gap sits.
Work the list from the most-cited page down, because in our 192-answer study "a nofollow mention inside a roundup that 24 answers cite is worth more here than a followed link on a page nothing retrieves."
What earns a spot on a list, and how do you pitch the owner?
No primary source says what makes an editor add a vendor to a list. Google, Microsoft, OpenAI, Perplexity, the FTC and the review platforms are all silent on it. So anchor your pitch on the standard that list owners are told to meet.
Google's guidance on writing reviews says: "When recommending something as the best overall or the best for a certain purpose, include why you consider it the best, with first-hand supporting evidence." It also asks writers to "Explain what sets something apart from its competitors." A good pitch hands the owner exactly that: the reason you belong on their list, and evidence they can check.
A pitch that works is short and specific:
- One plain sentence on who you are and who you serve. Our 2,895-answer study found the pages that most often went with a naming stated what we do and for whom. "Say who you are on the page, in one plain sentence." The list owner reads that sentence too.
- The specific fact that makes you belong on that list. Match it to the list's own purpose. For a "best for mid-market" roundup, say what you do for mid-market teams that the listed options do not.
- First-hand proof. A dated customer outcome the owner can verify, a public review profile, a product page that shows the feature. Link to it.
- An easy edit. Offer the one-line entry, written in their format, so adding you takes almost no work.
Send it to the person who maintains the page, by name. Skip the mass email. The owner is one person deciding whether you belong on their page, so write to that person about that page. If you hear nothing, follow up once, with something new (a fresh result, a new review, a changed feature). Then leave it.
Community threads work differently. Nobody maintains a Reddit thread for you to pitch. Reddit's rules ask users to "Participate authentically in communities where you have a personal interest, and do not spam or engage in disruptive behaviors", and not to "impersonate an individual or entity in a deceptive manner." Answer real questions in communities you already belong to, say that you work at the company, and let the thread decide.
How do review platforms like G2 and Capterra fit in?
They are lists with rules, and the rules are strict.
G2's community guidelines say "Anyone who is not an actual user of the product may not leave reviews." They add: "Employees working for the product's company and employees of a direct competitor may not leave reviews, as those would be considered biased." G2 says it "will never require or ask that only positive reviews be submitted to its site" and "will never suppress or otherwise mute/hide negative reviews", and it does "not edit or remove reviews at a seller's request".
Incentives are allowed on G2 but visible. "If a review is incentivized, G2 will clearly label the review as incentivized." The incentive is capped, and eligibility never depends on the opinion in the review. Reviews from business partners and guest users are labelled and do not count toward the G2 Score, so asking partners for reviews does not move the score.
Capterra's guidelines run along the same lines. Reviews must come from individuals whose identity can be verified and who are not affiliated with the vendor or a direct competitor, or financially interested in the product. Incentives must be offered equally to everyone eligible, regardless of the rating given. Vendors may not review their own product, post reviews on behalf of users, or commission or buy fake reviews, and they must ask a broad cross-section of users rather than picking people by expected sentiment.
Build a steady, honest review flow from real customers, and invite everyone rather than only your happiest accounts. Never post a review on a customer's behalf, which Capterra forbids, and never tie an incentive to a positive rating, which both platforms forbid. The FTC's 2024 rule separately bans fake reviews and buying reviews conditioned on a particular sentiment.
How much review platforms feed AI answers in your category is an open question. In our agency-prompt study they carried 0.5% of citations and G2 never appeared. That is one category on 8 prompts. Check your own prompts before you decide review platforms do or do not matter for you.
Can you pay for a placement on a best-of list?
No AI engine says either way. As our 192-answer study noted, "No engine publishes a policy on whether paid placement, press releases or directory listings change your citation odds." A claim that paid placement is rewarded or penalized in AI answers is a guess.
What does exist are search and advertising rules, and they apply whether or not an AI engine reads the page.
Google's link rules. Google's spam policies list "Buying or selling links for ranking purposes" as link spam, including "Exchanging money for links, or posts that contain links" and "Exchanging goods or services for links". The same policy allows paid links that are qualified with rel="sponsored" or rel="nofollow". Google's link guidance calls nofollow "still an acceptable way to flag them, though sponsored is preferred." So a paid slot with a followed link breaks Google's rules. A paid slot with a sponsored link does not.
Google's site reputation policy. Third-party content on a strong host site is not the issue in itself. It breaks the policy when it is published "mainly because of that host site's already-established ranking signals". Google's examples of content that does not break it include editorial articles and advertorial pages written to share content directly with the publication's readers. That is a Google Search risk for the host site. It says nothing about AI answers.
The FTC Endorsement Guides. The Guides (16 CFR Part 255, revised 2023) say a connection that might affect the weight or credibility of an endorsement, and that the audience would not reasonably expect, "must be disclosed clearly and conspicuously". A material connection "can include monetary payment or the provision of free or discounted products", as well as business, family or personal relationships, and the duty applies "regardless of whether the advertiser requires an endorsement in return". Intermediaries are covered too: "Advertising agencies, public relations firms, review brokers, reputation management companies, and other similar intermediaries may be liable for their roles in creating or disseminating endorsements containing representations that they know or should know are deceptive", and for endorsements that fail to disclose unexpected material connections. Paying through a third party does not move the disclosure duty elsewhere.
Swaps count. FTC staff, answering a question about two authors reviewing each other's books for free, said "the need to make a disclosure isn't limited to situations in which money changes hands." A "you list us, we list you" arrangement between two vendors raises the same disclosure question.
The 2024 fake-reviews rule. The FTC's final rule on fake reviews and testimonials, announced 14 August 2024, bans creating, buying or selling fake reviews, buying reviews conditioned on a particular sentiment, undisclosed insider reviews, and suppressing reviews with groundless threats. It lets the agency seek civil penalties against knowing violators.
Our own position is selectivity over volume. As the same study put it, "We do not buy hundreds of backlinks on sites that earn nothing from AI models." One disclosed placement on a list the engines actually read beats a batch of placements on pages nothing retrieves.
Should you publish your own best-of page?
Yes, as one piece of the plan, because it is the one list you control. Build it to the standard Google sets for ranked lists, and do not expect it to do the third-party job for you.
Google's reviews guidance asks for "first-hand supporting evidence" behind every "best" call, and says to "Ensure there is enough useful content in your ranked lists for them to stand on their own". It asks for "quantitative measurements about how something measures up in various categories of performance" and the "benefits and drawbacks of something, based on your own original research." Its people-first guidance adds two tests: "Is it self-evident to your visitors who authored your content?" and "Are you mainly summarizing what others have to say without adding much value?" A page that only re-lists names from other lists fails the second one.
The FTC rule has a line for this case as well. It "prohibits a business from misrepresenting that a website or entity it controls provides independent reviews or opinions about a category of products or services that includes its own products or services." Say who published the page, and do not present it as independent.
Format matters for AI answers. Microsoft says assistants like Copilot "break content down, a process called parsing, into smaller, structured pieces", and that "Bulleted lists, numbered steps, and comparison tables break complex details into clean, reusable segments." It also warns: "Don't hide important answers in tabs or expandable menus". Put the comparison table in the visible page.
Schema is less important than people think. Google says "There's also no special schema.org structured data that you need to add" for AI features. Its Review snippet markup covers a fixed set of types, including Software App. Its ItemList carousel covers only Course list, Movie, Recipe and Restaurant, so a B2B software list is not documented as eligible. And "The FAQ rich result feature is no longer shown in Google Search results". Microsoft, for its part, says "Schema can label your content as a product, review, FAQ, or event". Use markup that matches the visible text, and do not expect it to win you a slot.
Your own page still cannot do the whole job. In our 2,895-answer study, the pages that said plainly who we are and who we serve were named far more often than our broad list, which the study sums up as "a page where you are one of twenty entries does less for you than a page about you." Your own list helps. A clear page about you helps more. A place on the lists the engines already read covers the prompts where neither of those gets retrieved.
How do you check you are in the lists that count?
Ask the same questions, on each engine, on a schedule, and record two things per answer: which pages it read, and which brands it named.
Measure namings per engine, on the same questions, every week, as our 2,895-answer study recommends. Keep reading and naming as separate numbers. In that study, the share of reads that turned into a naming ran from 41.5% on Perplexity to 77.5% on Google AI Overviews. A single blended number would hide which engine is where your gap sits.
For each target list, track three states over time:
- Not retrieved. The engines do not read the list for your prompts. Deprioritize it.
- Retrieved, you are absent. The list is read and you are not on it. This is your pitch queue.
- Retrieved, you are on it. Check whether the answers that read it now name you, engine by engine.
Open the live list yourself when your tracking says you are on it, and search the page for your name. Blocked pages can look read in a tracker and return a challenge page when you visit.
Set expectations on speed. Our guide to getting named in AI search found a well-structured page on a brand with some existing authority can get named within a day, while "climbing to a dominant share of answer on a competitive prompt is a months-long compounding effort." A new list placement sits in between: the engine has to retrieve the updated page before it can name you from it.
To check where your brand stands across ChatGPT, Claude, Gemini and Perplexity against your top competitors, start with an AI visibility audit. For placement in the rosters, reviews and communities each engine retrieves from, see our GEO agency service.
