KnowBe4 was the only security awareness training tool that ChatGPT and Gemini named in all 6 of the 6 buyer questions we ran, a 100% hit rate no other product came close to.
On 2026-08-07 we asked ChatGPT and Gemini six versions of the same shopping question, from the plain "best security awareness training software" to "best free" and "most affordable," let each assistant run one live web search, and logged every product it named and every source domain it cited. Across the six answers, 21 different products came up. Only KnowBe4 appeared in every one.
After KnowBe4, the field splits fast. Hoxhunt was the clear second at 4 of 6 answers (67%), and it is the only other product above the halfway line. Right behind it sits a five-way tie at 3 of 6 (50%): Cofense, Infosec IQ, NINJIO, Proofpoint and Wizer. So there is one near-universal answer and then a crowded middle with no obvious runner-up.
The tail is long and thin. 11 of the 21 named products showed up in just a single one of the six answers (one appearance each, 17%). SoSafe, SafetyCulture, EasyLlama, Symbol Security, CyberHoot, MetaCompliance, Barracuda Security Awareness Training and Mimecast Engage were each named for exactly one question and absent from the other five.
The shortlist changed with the question
The most useful thing in this data is that the six questions did not return the same shortlist. Change the wording and the names change, and that is the whole story of how AI answers a category.
The bare "best security awareness training software" returned 6 names: KnowBe4, Proofpoint, Infosec IQ, NINJIO, Hoxhunt and Mimecast Engage. Mimecast Engage showed up here and nowhere else, its only appearance in the entire run.
Add one word, "free," and the answer swaps to a different 8-name set: Wizer (named first, ahead of KnowBe4), KnowBe4, Proofpoint Security Awareness Training, Hoxhunt, CyberHoot, Cofense, Mimecast Awareness Training and Barracuda Security Awareness Training. Three of those, CyberHoot, Mimecast Awareness Training and Barracuda Security Awareness Training, appeared only for the free question. The free query even flipped the leader: Wizer topped the free list but never appears for the plain "best" question.
Ask for "most affordable" and you get the shortest answer of all, just 4 names: KnowBe4, Wizer, Hoxhunt and Cofense. The affordable set drops nearly every enterprise brand (no Proofpoint, no Infosec IQ, no NINJIO) and leans on the budget trio of Wizer, Hoxhunt and Cofense behind KnowBe4.
The other phrasings surfaced their own one-off names. "For small teams" pulled in EasyLlama and Symbol Security, each named only there. "What should I use" uniquely surfaced SafetyCulture and SoSafe. "Recommendations" uniquely surfaced Cofense PhishMe, Security Compass and MetaCompliance. Only KnowBe4 rode through all six phrasings unchanged. For everyone else, "does AI recommend us" has no single answer: it depends on whether the buyer types best, free, affordable or for small teams.
The source map: who fed the answers
Reddit was cited in all 6 of 6 answers, more than any other community source, and it was the only community site to appear at all. No Quora thread showed up; Reddit was the entire community-source story behind every answer we got.
Two other domains also hit 6 of 6: knowbe4.com and securitycompass.com. Alongside Reddit, those three fed every single answer. Gartner was next at 5 of 6 (83%), and SoftwareReviews appeared in 2 of 6 (33%).
Here is the key insight for anyone selling in this category: KnowBe4's own website, knowbe4.com, was cited in all 6 of 6 answers, and the KnowBe4 product was named in all 6. The vendor that published the most is the vendor that got recommended the most. Hoxhunt shows the same shape smaller: hoxhunt.com fed 3 answers and Hoxhunt was named in 4. Wizer, Proofpoint, SoSafe, Symbol Security, EasyLlama and Adaptive Security (adaptivesecurity.com, cited 3 times, product named twice) were all named off their own domains too.
The counter-example is just as instructive. securitycompass.com was also cited in all 6 answers, tied with knowbe4.com, yet its own product was named far less: Security Compass once and Security Compass Application Security Training twice. Its pages read as roundups that list the whole field, so it earns citations without turning them into mentions of itself. Being cited as a source and getting your own product named are two different wins.
And the review directories everyone assumes own this category barely featured. G2 appeared in 1 of 6 answers (17%) and GetApp in 1 of 6 (17%). Capterra did not appear at all. In this run the directories were the least-used sources, far behind Reddit (6 of 6), Gartner (5 of 6) and the vendor roundups.
What you would do to get named here
The lever is the source list, not your market share. The assistants named whoever was written about on the pages they retrieved for each question, nothing more. EasyLlama and Symbol Security prove it: neither is a category giant, both were named exactly once, and both rode a single citation of their own domain that happened to get pulled for the "small teams" question. Being on the retrieved page beat being the market leader.
So the practical move is to be in the exact pages these answers draw from. That means the Reddit threads behind all 6 answers, the Gartner and SoftwareReviews roundups (5 of 6 and 2 of 6), and the published "best security awareness training" listicles on domains like securitycompass.com and knowbe4.com that were cited 6 times each. Publishing your own well-sourced comparison is a lever too: KnowBe4's own domain fed every one of the 6 answers.
Then pick the query you actually want to win. If you are the affordable option, the four-name "most affordable" answer (KnowBe4, Wizer, Hoxhunt, Cofense) is a much shorter list to break into than the crowded "best" set, but only if you are on the pages that answer the free and affordable questions (wizer-training.com, getapp.com, usecure.io, astound.com), which are not the same pages that answer the plain "best."
The full list, counted
| Product | Named in | Share |
|---|---|---|
| KnowBe4 | 6 of 6 | 100% |
| Hoxhunt | 4 of 6 | 67% |
| Cofense | 3 of 6 | 50% |
| Infosec IQ | 3 of 6 | 50% |
| NINJIO | 3 of 6 | 50% |
| Proofpoint | 3 of 6 | 50% |
| Wizer | 3 of 6 | 50% |
| Adaptive Security | 2 of 6 | 33% |
| Proofpoint Security Awareness Training | 2 of 6 | 33% |
| Security Compass Application Security Training | 2 of 6 | 33% |
A note on the table: the assistants named some brands under two labels, and we kept both rows rather than summing them into a number the data never showed. Proofpoint was counted as both "Proofpoint" (3 of 6) and "Proofpoint Security Awareness Training" (2 of 6). Security Compass split into "Security Compass" (1 of 6) and "Security Compass Application Security Training" (2 of 6). Cofense appeared as "Cofense" (3 of 6) and "Cofense PhishMe" (1 of 6), and Mimecast as "Mimecast Engage" (1 of 6) and "Mimecast Awareness Training" (1 of 6). When AI cannot keep your product's name straight, that fragmentation is its own thing worth tracking.
Where Glotier fits, and where it does not
Glotier does not sell security awareness training software, so we are correctly absent from all 6 of these answers, and we would be suspicious of our own tool if we were not. This is not our category and we are not pretending it is.
But this page is exactly what Glotier runs for a customer's own category: the same six-question fan-out, the same count of who gets named out of six, the same source map showing which Reddit threads, roundups and vendor pages the answer actually came from. If you sell security awareness training, the version of this page that matters is the one with your name in the table and your competitors' counts next to it.
The check is free, takes about a minute, and needs no account. Run your own category and see whether ChatGPT and Gemini name you, and which sources decide it.