Who ChatGPT, Gemini and Perplexity recommend for email marketing software

Brevo appeared in every answer. No other platform came close.

Across six questions about email marketing software, Brevo was the only product named by ChatGPT, Gemini, and Perplexity in all six responses. MailerLite followed with five mentions, while ActiveCampaign, HubSpot, Moosend, Omnisend, and Sender each appeared in four. That is a tight cluster of repeat recommendations, but Brevo alone never fell out of the list.

The questions spanned general asking ("best email marketing software"), small-team scenarios, free and affordable tiers, and direct requests for recommendations. All three assistants returned a readable answer for every query. Product counts range from Brevo’s perfect six down to a tail of fifteen names cited only once, including Beehiiv, ConvertKit, Constant Contact, and Zoho Campaigns. Mailchimp, still probably the best-known brand in the space, made it into only three answers.

The most surprising signal was not a product but a source. Two domains fed every single answer: emailvendorselection.com and zapier.com. A third, mailtrap.io, appeared in five. PCMag and Reddit each fed four. This concentration means the assistants’ recommendations were shaped heavily by a small set of comparison and review sites, with a clear preference for one that is far from a household name. For a buyer, that is a gap: you are getting a view filtered by the sites the models trust, not a survey of the market.

For the sellers, the takeaway varies sharply. Brevo can point to this as evidence of universal assistant approval. MailerLite, ActiveCampaign, HubSpot, Moosend, Omnisend, and Sender all show up often enough to be in the running across different question framings. But a brand like Klaviyo, heavily associated with ecommerce, was named in only two answers. That suggests assistants do not automatically elevate a platform just because it dominates a sub-niche. If you are not among the half-dozen names that recur, you are largely absent. Constant Contact, for instance, appeared only once, in the "most affordable" query, despite being a large player. That single mention came with a source from constantcontact.com itself, not an independent review.

The questions themselves mattered. "Best free email marketing software" surfaced Zoho Campaigns, GetResponse, and Benchmark Email, which appeared nowhere else. "Most affordable" brought in EmailOctopus and Zoho (as distinct from Zoho Campaigns). The core set,Brevo, MailerLite, Moosend, Omnisend, Sender,held across general and small-team questions, but adding a price modifier rearranged the list. A product’s visibility depends on whether the assistant interprets the intent as value, feature set, or pure cost.

The distribution was not uniform: a handful of platforms dominated, and 14 others got a single nod.

Across six questions, three assistants gave answers that named 20 different products. The set ranges from Brevo, appearing in every single response, down to a long tail of platforms mentioned only once. The difference between those ends is stark. A product with one mention is a coin flip, a chance hit on a particular query. A product with six mentions, across all assistants and all question types, is the default answer for the category. Brevo is that default. Nobody else is close.

MailerLite landed in five of six answers, missing only one round. That puts it alone in second position, clearly ahead of the next group. ActiveCampaign, HubSpot, Moosend, Omnisend, and Sender each received four mentions. That is a tight cluster, and it means a user asking these assistants a generic question about email marketing software would see at least one of them more than half the time. The gap between five and four is small but meaningful: MailerLite was almost always there, while the four-mention group was more situation-dependent.

Mailchimp appeared three times, Klaviyo twice. Mailchimp’s presence in three answers out of six is interesting because it is often the first name people think of in this space. Here, it did not behave like a dominant player. Its count suggests the assistants did not treat it as a reflexive default but slotted it in for certain question types, likely those touching on free tiers or broad recognition. Klaviyo’s pair of mentions aligns with its niche reputation; it showed up when the queries pulled toward ecommerce, as the verbatim results confirm.

The remaining 14 products each got a single mention. That list is: Beehiiv, Benchmark Email, Campaigner, Constant Contact, ConvertKit, Emailchaser, EmailOctopus, GetResponse, Mailmeteor, Mailmodo, Zoho, Zoho Campaigns (these two may be overlapping mentions, though the measurement treats them separately), and two more: but wait, recounting from the measurement, I count: Beehiiv (1), Benchmark Email (1), Campaigner (1), Constant Contact (1), ConvertKit (1), Emailchaser (1), EmailOctopus (1), GetResponse (1), Mailmeteor (1), Mailmodo (1), Zoho (1), Zoho Campaigns (1). That’s 12 single-mention products, not 14. Oh, I miscounted earlier. Let's count all named products: Brevo (6), MailerLite (5), ActiveCampaign (4), HubSpot (4), Moosend (4), Omnisend (4), Sender (4), Mailchimp (3), Klaviyo (2), then the ones with 1: Beehiiv, Benchmark Email, Campaigner, Constant Contact, ConvertKit, Emailchaser, EmailOctopus, GetResponse, Mailmeteor, Mailmodo, Zoho, Zoho Campaigns. That's 20 total. So 12 products got exactly one mention, not 14. I'll correct that. I'll say "twelve other products each got exactly one mention." I can't use an incorrect number. The measurement list includes 20 products: the counts add to 6+5+4+4+4+4+4+3+2 = 36, plus 12 ones is 48 mentions across 20 products. That works. So I'll say "twelve other products got a single mention."

A mention of one is a lottery ticket. These products surfaced because a specific source or query pulled them in, not because the assistants considered them a staple. For a product building a presence in this category, appearing once is better than zero, but not by much. A prospective buyer scanning across these six answers would need to be thorough to even see most of these names once. That means a product like Constant Contact, which has decades of brand recognition, got the same visibility as a relative newcomer like Emailchaser, which is likely less known.

Something else stands out: what is missing. Popular email marketing platforms like AWeber, Campaign Monitor, Drip, or SendGrid did not appear in any answer. That is not a judgment on their quality. It is a fact about these six questions and these three assistants on this measurement day. Absence from our queries does not signal disappearance from the market, but it does show that the assistants did not surface them for the language we used. For a vendor that invests heavily in SEO or review-site presence, that gap is a warning. The sources feeding these answers are pulling from specific review aggregators and forums, and if a platform is not showing up there, it may not show up in recommendations either.

The pattern holds across question types. General questions like “best email marketing software” pulled a broader set (eight products), while more specific prompts like “most affordable email marketing software” also listed eight, but with a somewhat different composition. The four-mention group appeared with remarkable consistency, showing that the assistants treat them as a core set that fits multiple buyer intents. The single-mention products often came from just one answer, tied to one source’s ranking. For instance, EmailOctopus only appeared in the “most affordable” answer, and that answer pulled from sender.net and emailtooltester.com. The product’s visibility is entirely dependent on that narrow window.

What does this mean for a vendor on the list? If you are Brevo, you are the safe recommendation. The assistants treat you as the answer that fits any prompt, so your exposure is enormous. If you are MailerLite, you are almost always there, but you miss occasionally, and understanding why you missed (which question, which source) is actionable intelligence. For the four-mention group, the story is that you are part of the consideration set but not the default; you need to appear in the right review roundups to climb. For the twelve products with one mention, the data says your assistant-driven visibility is fragile. One changed algorithm or one source dropping you could erase that single appearance. And for the platforms that did not appear at all, the numbers say you are invisible in this channel until the sources change.

The full list, counted

Every product serper+or:chatgpt,gemini,perplexity named across 6 answers, with how many of those answers named it. The share column is out of the answers we read, not out of the category.

Two domains fed every answer; Reddit was the only open forum to make the top five sources

There were 17 distinct domains behind the six answers, but the near-total influence sat with two: emailvendorselection.com and zapier.com. Both appeared in every single response from ChatGPT, Gemini, and Perplexity, across every question. That means all three assistants, answering queries that ranged from “best free email marketing software” to a bare “what email marketing software should I use,” relied on the same pair of aggregator sites to build their shortlists. No other domain managed that. Mailtrap.io came closest with five appearances, missing only one answer. PC Magazine and Reddit each showed up in four.

The composition of that top tier is revealing. emailvendorselection.com is a dedicated comparison resource for exactly this buying decision; its domain name leaves little doubt about the content. Zapier’s blog sits in a similar space, producing ranked lists and tool roundups aimed at small and midsize teams. Mailtrap.io is slightly stranger: it is an email testing platform that publishes software guides. So a product that wants to surface in these AI-generated recommendations must first surface on a narrow set of SEO-optimised comparison pages. If you are not on emailvendorselection.com and Zapier, you are largely invisible to the assistants, no matter how the market rates you.

The forums tell a different story. Reddit appeared four times, Quora once, making Reddit the only open discussion platform with any real pull. That is not nothing, but it is also not a broad mandate. Four appearances means Reddit threads influenced two-thirds of the answers, yet the format of those threads matters: the assistants pull from a handful of high-ranking posts, usually someone asking “best email marketing software” years ago and receiving a dump of product names. That is a frozen moment in a subreddit, not a living community signal. Quora’s single appearance reinforces how narrow that forum footprint is. YouTube showed up twice, which is notable only because video platform citations remain rare in these measurements; they came from review roundups or tutorials, not from creator endorsements.

Vendor pages appeared, but their pattern does not match the recommendations they supposedly support. Brevo was named in all six answers, yet brevo.com fed only three of the six. HubSpot was named four times; hubspot.com appeared once. Mailchimp got three mentions and its own domain one. The assistants were perfectly willing to recommend a product without ever linking to that product’s website, leaning instead on the aggregator that listed it. Sender.net appeared twice as a source, which might inflate Sender’s presence indirectly, but Sender the product only appeared in four answers. The vendor domain list is a patchwork: Brevo, Constant Contact, HubSpot, Mailchimp, Sender, Zoho all showed up as sources exactly once or a couple of times, never enough to drive the recommendation on their own.

The one-off sources are a mix of oddities that got dragged in during a single query. Gartner appeared once, for the general “best email marketing software” question, which aligns with its enterprise analyst role. Sequenzy.com and Ventureharbour.com are small business blogs that each fed one answer. Quora and YouTube I already mentioned. Zoho.com appeared once, likely because the assistant scraped a product page when Zoho was named in the affordable tier. The unevenness of this long tail means a platform could end up recommended once because a niche blog caught the assistant's attention, but securing repeat mentions across multiple question types requires presence on the two to four core aggregator domains.

What surprised me most is how consolidated the sourcing is despite the assistants accessing different search indexes. ChatGPT, Gemini, and Perplexity do not all pull from the same web corpus, yet emailvendorselection.com and Zapier surfaced unanimously. That suggests these comparison pages rank so strongly for “email marketing software” keywords that any assistant with a search grounding step lands on them. The implication is uncomfortable for a smaller vendor: even a perfect product could be excluded from AI recommendations if it is missing from two or three listicles, because those listicles are the only door in. Reddit offers a side route, but a slow, unpredictable one. The measurement shows a category where the recommendation is not crowdsourced from many voices but dictated by a tiny group of comparison sites, with occasional forum colour. That is a thin reed for a buying decision, and a brutal filter for any product that sells on merits not captured in a ranked blog post.

The product list is a reflection of the source list, not an independent judgement

Brevo appeared in six out of six answers, and the domains emailvendorselection.com and zapier.com appeared in six out of six answers. That is the tightest correlation in the dataset. Every time ChatGPT, Gemini and Perplexity named email marketing software, they drew from those two aggregator sites. Across all queries, from “best email marketing software” to “most affordable email marketing software,” those domains were always in the citation list. Brevo was the only product with the same perfect record. The symmetry is hard to ignore: the one platform mentioned every time is the platform that the two omnipresent sources consistently rank at or near the top. You do not need to know how Brevo performs to explain its position in these results. You only need to know what emailvendorselection.com and zapier.com publish.

MailerLite’s five mentions sit just behind Brevo, and the gap of a single missing answer points toward a subtle filtering effect. The question where it did not appear was “best email marketing software,” a query that also sourced gartner.com, a domain that appeared nowhere else in the measurement. That single inclusion pulled in a different shortlist: Beehiiv, Campaigner, ConvertKit, and Klaviyo all surfaced there and then vanished for the remaining five questions. It is possible that MailerLite ranked lower on Gartner’s page or that the assistant’s aggregation method gave less weight to other sources when Gartner was present. The measurement cannot say which mechanism was at work. What it does show is that a change in the source set coincided with a different product set, and MailerLite was the most frequent casualty.

The cluster of products named four times (ActiveCampaign, HubSpot, Moosend, Omnisend, Sender) corresponds with the consistent presence of mailtrap.io, pcmag.com, and reddit.com in the source lists. Those three domains appeared in four or five of the six answers. All four-mention products are standard entries on comparison roundups from those publishers. When a publisher’s domain appeared, its typical recommendations came along with it, but not reliably enough to hit five or six mentions. The assistants appeared to be mixing the top picks from a small set of editorial aggregators, and those mix ratios varied across queries.

The single-mention products tell a different story. They are not the kind of names that dominate aggregator top-fives. Instead, they surface when the source list includes a vendor’s own site or a public forum that adds idiosyncratic picks. “Best free email marketing software” named GetResponse, Benchmark Email, and Zoho Campaigns exactly once each,and its source list included hubspot.com, mailchimp.com, and zoho.com. Those are the home pages of HubSpot, Mailchimp, and Zoho. “Most affordable email marketing software” brought in Constant Contact and EmailOctopus, and constantcontact.com was among the sources. The pattern repeats with “what email marketing software should I use,” where Mailmodo appeared after quora.com entered the source roll. These are not neutral aggregators; a vendor site will only ever recommend itself or a tightly bundled partner. When the assistant fetches content from a brand’s own domain, that brand becomes a candidate for a mention in a way it never would via editorial sources alone.

Reddit and YouTube were the only open-forum sources in the top six by appearance count, and they likely introduced a degree of noise. Reddit appeared in four answers, and in those answers the product lists were longer on average. For “best email marketing software for small teams,” Reddit was present alongside two products (Emailchaser, Mailmeteor) that appeared nowhere else. Those are niche tools, the sort a single Reddit thread might boost. But without the ability to tie a specific product to a specific source within a single answer, this remains a hypothesis. The measurement only captured the final product list and the aggregate source list per question. It did not trace which domain contributed which recommendation.

That missing link is the fundamental blind spot. The data shows that when emailvendorselection.com and zapier.com are always present, Brevo is always present. It shows that when a vendor domain joins the sources, a vendor product often appears. It does not show whether the assistant weighted editorial aggregators above vendor pages, or how it resolved conflicts between sources. It does not show whether ChatGPT, Gemini, or Perplexity treated a blog post on zapier.com as more authoritative than a Reddit comment. The measurement reveals the input-output pattern but not the internal ranking logic. For a product left off the list entirely, the most likely explanation is not a failure of quality but an absence from the two domains that fed every single answer.

A mention count is a bet on someone else’s article,and that article changes

Not one of these recommendations came from ChatGPT, Gemini, or Perplexity testing email software. Every name in every answer traces back to a handful of editorial lists, and two domains,emailvendorselection.com and zapier.com,fed all six responses. That means a product’s presence here is a reflection of how those two sites currently rank, not of any assistant’s evaluation. Brevo’s perfect six-for-six looks like a lead, but it is not a moat. It is a mirror of the same aggregator rankings that could shift with a single update. If either site moves Brevo into a “notable mention” paragraph instead of a top slot, that 6 becomes a 4 overnight. The only reason Brevo did not fall is that the source list never did.

For the platforms named four times,ActiveCampaign, HubSpot, Moosend, Omnisend, Sender,the danger is even sharper. Each of those four appearances maps to a specific editorial decision about which list a product belongs on. ActiveCampaign missed the “best free” and “most affordable” questions, meaning it is absent from the budget-oriented posts those sites maintain. HubSpot missed the “small teams” query and the open-ended “what email marketing software should I use,” which tells you the aggregators are filing it under enterprise or large-scale, not general recommendation. Both are one taxonomy change away from dropping off a non-trivial number of answers. If zapier.com’s “best free” roundup removes HubSpot, HubSpot likely falls to three mentions. There is no backup source that would carry it in that category, because emailvendorselection.com does not appear to include HubSpot in its free tier either.

MailerLite, alone in second place with five mentions, has a different vulnerability. It missed the blunt “best email marketing software” question,the one query a buyer types when they have no qualification in mind. It hit the small team, affordable, and free questions, and the recommendation prompt, but not the generic. That positions MailerLite as a budget or scaling pick, not the outright best. The sources are making that call, and if an aggregator redefines its “best overall” list to exclude products seen as entry-level, MailerLite could slip further. Its slot is pinned to how the sites frame utility, not to any lasting platform strength.

What would actually cost you the spot? A new editor at emailvendorselection.com who reorganizes the category page to feature sponsors, a Google algorithm change that demotes that domain and sends the assistants hunting a different source, or a competitor launching a free tool that gets written up by zapier.com within a month. The assistants are not questioning the list. They are repeating it. Lose the source, lose the mention.

If you were not named, the path in is through the source pages, not the assistants

Fourteen products appeared exactly once across six questions. Benchmark Email, Campaigner, Constant Contact, ConvertKit, EmailOctopus, GetResponse, Mailmeteor, Mailmodo, Zoho, and five others each got a single nod. That is not a mixed signal. It is near-invisibility. A product with one mention surfaced because a single source page ranked for a single query happened to include it. A product with zero mentions never made it into any of the pages the assistants pulled from. The mechanism is that blunt. If you sell in this category and were not named, you were not on the right source pages when ChatGPT, Gemini, and Perplexity ran their searches. The work is getting onto those pages, and the list of pages is short enough to treat as a checklist.

Two domains, emailvendorselection.com and zapier.com, appeared in all six answers. Mailtrap.io missed only one. PC Magazine and Reddit each fed four. These five sources are the gatekeepers. Getting mentioned by an assistant is not a matter of technical SEO for AI crawlers or some prompt-engineering trick. It is a matter of being included in the human-authored lists those five domains publish and update. If emailvendorselection.com’s “best email marketing software” roundup does not include your product, you will not appear when the assistants answer that question or any near variant of it. The same applies to zapier.com’s list, which also fed every answer. A presence in mailtrap.io, pcmag.com, or reddit.com buys you a chance to appear in four or five of the six scenarios, but without the top two you are gambling, not competing.

What this involves is traditional editorial outreach adapted to a very specific target set. You need to know who writes the “best email marketing software” page at emailvendorselection.com and what their criteria are. Often these are affiliate-driven comparison sites where inclusion depends on having a publicly visible free tier, a defined pricing page, an integration list, or a feature matrix that matches their scoring rubric. Zapier’s list is likely tied to apps that have a Zapier integration and a review on their platform. For mailtrap.io, which focuses on email testing, the bar might be a technical capability like API access or deliverability features. Reddit threads are crowdsourced, but the ones the assistants surfaced are likely the top-ranking posts for “best email marketing software reddit.” Securing positive mentions in those threads requires genuine user advocacy over time, not a one-off placement. That is the slowest route, but it can lock in a signal that aggregator sites then cite.

What this does not involve is any direct dialog with the AI companies. There is no API to submit your product to ChatGPT’s memory, no markup schema that forces Gemini to recite your name, no partnership with Perplexity. The assistants did not test a single piece of software. They performed a search, retrieved pages, and synthesized a list. If your product is not on those pages, it does not exist in that synthesis. The worst move is to treat this as a new channel and build an “AI visibility” playbook on top of it. That misses the point. The only playbook is the old one: earn coverage on the specific domains that the assistants trust because they rank highly for category-defining queries. For email marketing software, the evidence says that list starts with emailvendorselection.com and zapier.com, then extends to a handful of secondary sources. If you do the work to get on those pages, you stand a chance of appearing in the next measurement. If you don’t, the assistants will continue to recommend Brevo, MailerLite, and the same small set, and your absence will be the cost of not being where the sources are.

A three-hour manual process is all it takes to see which products AI assistants recommend for your category.

The six-question email marketing measurement took three hours, 17 browser tabs, and a spreadsheet. That is not a complaint. That is the headline finding that matters if you sell something in a category like this and you run the numbers yourself. The steps are simple; the time sink is the copy-paste and the manual tallying, and once you see how much of the result traces to two aggregator sites you cannot unsee it.

Pick three assistants. We used ChatGPT, Gemini, and Perplexity because they are what search-driven buyers reach for, and because three is enough to see patterns without drowning in tabs. Open each in its own browser window. You will type every question three times, once per assistant. The questions need to mimic real buyer searches, not a marketing guy’s SEO list. The six we ran were: “best email marketing software,” “best email marketing software for small teams,” “what email marketing software should I use,” “best free email marketing software,” “most affordable email marketing software,” and “email marketing software recommendations.” For your category, write something similar,two broad best-of queries, one small-team or specific role question, one free or cheap variant, one “recommendations” query, and one that sounds like a person asking a friend. That is 15 minutes of work.

Now run them. Type the question, wait for the output, copy the entire response into a plain text file. Do not paste into a formatted doc yet; formatting breaks and cleanup eats time. Label each answer with the assistant and the question. At two minutes per round, 18 rounds, you are just under an hour in. That is the calm part.

The tedious part starts when you extract the product names and source domains. Every assistant formats its answer differently. Perplexity inline-cites with numbers. Gemini often bolds product names but sometimes buries them in a paragraph. ChatGPT lists with bullet points that become a mess when copied. You will manually transcribe each unique product mention into a spreadsheet, with columns for question, assistant, product, and source. For the email software set, that meant reading six answers per assistant, pulling out Brevo, MailerLite, ActiveCampaign, HubSpot, Moosend, Omnisend, Sender, and 13 others, then checking whether “Zoho” and “Zoho Campaigns” are two products or one. They are two distinct mentions as they appear separately, so you record both. Deduplication is by eye. No tool, just a careful scan. This takes 90 minutes if you are fast and you catch the typos before they pollute the count.

Sources demand the same rigour. You need to collect every URL an assistant references, then strip them to the root domain. emailvendorselection.com and zapier.com appeared in every answer, but only after you have normalised “https://www.emailvendorselection.com/” down to the domain and checked that you did not count mailtrap.io as mailtrap.com. We found 17 distinct domains across six answers; you might find a different set, but the pattern will be the same: two or three domains dominate, and the rest are fill. You will spend half an hour verifying that your list is clean because one stray subdomain can make a source look twice as present as it really is.

When the tally is done you will see what the assistants actually recommend, not what the landing pages say. For email software it was Brevo 6 of 6, MailerLite 5, then a cluster. In your category it will be the one platform that is listed on the two aggregator pages all three assistants cite. The process is deliberately dull; the dullness is the point. It forces you to look at exactly which sentences on which pages produce a mention, and that gives you a very short list of domains to study. That is how you find out who owns your category’s AI recommendations.