Does being mentioned more by AI make a tool more likely to be recommended first?

On average yes, and for individual tools often not. Across the 84 tools mentioned at least 6 times the correlation between mention count and first-rate is 0.45 (critical value about 0.22 at this sample size). The most-mentioned quartile averages a 34% first-rate against 7% for the least-mentioned. But 47 of those 84 tools have a first-rate of exactly zero, including Sprout Social at 17 mentions across 8 engines.

What does "named first" mean in this study?

The first tool listed in the engine's answer. It is not a separate ranking signal: on all 288 recorded answers the #1 was literally the first name in the response. It matters because list position is what a reader sees before deciding to keep reading.

Which tools are mentioned most but never named first?

Sprout Social (17 mentions, 8 engines), Copy.ai (15 mentions, 8 engines), SEMrush (15 mentions, 5 engines), Instapage (14 mentions, 7 engines), Google Forms (13 mentions, 7 engines). Each appears repeatedly across multiple independent engines and never occupies the first position in any of them.

Is a mention worth anything if it never converts to first place?

Yes, and less than it is usually sold for. Appearing in an answer at all puts a tool in front of a buyer who asked a purchase question. What this data shows is that appearing more often does not reliably move a tool up the list, so visibility spend justified by "we will get you mentioned" is buying presence rather than position.

How was this measured?

8 AI answer engines were asked the same buying questions across 16 software categories, and every tool named in each answer was recorded in order. Captured names are filtered to remove prose fragments before counting. Tools mentioned fewer than 6 times are excluded from rate calculations, because a zero out of two is not evidence.

61% of AI Tool Mentions Go to Tools That Are Never Named First

By Vincent Couey. Published August 27, 2026. Original research · 9 min read

The GEO industry sells mentions. This measures what a mention is worth. Across 288 recorded answers from 8 AI engines, 441 distinct tools were named. 364 of them were never once named first, and those tools hold 61% of every mention in the corpus.

Restrict it to tools with a real sample, mentioned 6 times or more, and 47 of 84 still have a first-rate of zero. Sprout Social appears in 17 answers across 8 separate engines and leads none of them.

61%

of all mentions go to tools never named first

of 84 well-sampled tools have a first-rate of zero

0.45

correlation between mentions and first-rate (critical 0.22)

of 16 categories where the most-mentioned tool is not the one named first most

Mentioned everywhere, first nowhere

These are the tools an AI visibility report would show as healthy. They appear in answer after answer, across engines that share no infrastructure and disagree with each other about almost everything. Not one of them has ever been the first name in an answer we recorded.

Mentions, among tools with a first-position count of exactly zero

Sprout Social17

CACopy.ai15

SEMrush15

Instapage14

GFGoogle Forms13

ZCZoho CRM12

CConvertKit12

Clearscope12

LSLinkedIn Sales Navigator11

JJira11

ZoomInfo11

Pipedrive10

Bars are grey because the quantity they encode is the one that did not convert. 47 tools clear the 6-mention floor with zero firsts; the 35 not shown continue down to 6 mentions.

And the tools that convert almost every mention

First-rate: share of a tool's mentions where it was named first

Calendly88%

Zendesk71%

Ahrefs69%

Jasper67%

DDialogflow63%

Apollo.io61%

SurveyMonkey60%

Unbounce59%

Google Analytics57%

Hootsuite53%

HubSpot52%

Asana50%

Same corpus, same 6-mention floor, opposite behaviour. Calendly converts 88% of its 16 mentions into first position. The gap between this figure and the one above is not a difference in how much has been written about these tools; several of the zero-rate tools above are mentioned more often than several of the tools here.

Does volume buy position? Partly, and not reliably

It would make a cleaner article to say that mentions and position are unrelated. They are related. Across the 84 tools with at least 6 mentions the correlation is 0.45, comfortably past the 0.22 that would be needed to call it real at this sample size. Sorted into quartiles by mention count, first-rate rises with volume, though the bottom two quartiles are indistinguishable from each other.

Mean first-rate by mention-count quartile

6 to 7 mentions7%

7 to 9 mentions7%

9 to 14 mentions19%

14 to 44 mentions34%

The most-mentioned quartile averages 34%, the least-mentioned 7%. And within that top quartile, 4 of 21 tools still convert nothing at all. The aggregate relationship is real and it does not describe the individual case.

No single engine is responsible for this

A zero first-rate would be much less interesting if it came from one or two narrow engines. It does not. All 8 of the 8 engines measured name tools from the zero-rate set, and none of them promotes one to the top of an answer.

Zero-rate tools named, by engine

Cohere Command-A42

GPT-OSS 120B42

Qwen 3.640

Llama 4 Scout28

Gemini 2.5 Flash-Lite25

Perplexity25

ChatGPT19

Llama 3.3 70B via Groq16

Counts range from 16 to 42, which tracks how many tools each engine names overall rather than any difference in behaviour: as a share of each engine's own named field the range is 18% to 28%. 3 of these tools are named by every engine in the study and led an answer on none of them: Sprout Social, Copy.ai, Jira.

That matters for what a vendor can do about it. A pattern confined to one model is an integration problem with a workaround. A pattern that holds across 8 engines built by different companies on different corpora is reflecting something upstream of all of them.

Presence and position are two different purchases

Presence

441 tools

Named at least once by at least one engine. This is what "get mentioned by AI" means, and it is a crowded room: 288 answers named an average of 5.4 tools each.

Position

77 tools

Ever occupied the first slot. 17% of the named field. Everything a reader sees before deciding whether to keep reading happens here.

Same corpus, same answers. The two columns are not stages of one funnel: 47 of the most-mentioned tools sit permanently in the left one.

The 5 categories where the most-mentioned tool is not the one named first

Aggregated across a whole corpus the gap is easy to dismiss as noise. Broken out by category it has names attached. In 5 of 16 categories the tool that appears in the most answers is not the tool most often placed at the top of them.

Category	Mentioned most	Named first most	Answers
Social media management	Buffer16 mentions	Hootsuite8 firsts	18
Survey / forms	Typeform16 mentions	SurveyMonkey6 firsts	18
AI copywriting	CACopy.ai14 mentions	Jasper10 firsts	18
Product analytics	Mixpanel14 mentions	Amplitude6 firsts	18
CRM	ZCZoho CRM12 mentions	Salesforce6 firsts	18

In the other 11 categories the two columns agree, which is the case most people assume is universal. It is not, and the exceptions are not obscure tools: they are category incumbents whose visibility is excellent and whose position is not.

Every well-sampled tool, mentions against firsts

Tool	Mentions vs first positions	First-rate	Engines
HubSpot	4423	52%	7
Mailchimp	3110	32%	8
Ahrefs	2618	69%	8
Klaviyo	243	13%	8
ActiveCampaign	243	13%	8
Apollo.io	1811	61%	6
Unbounce	1710	59%	8
Sprout Social	170	0%	8
Calendly	1614	88%	8
Buffer	166	38%	8
Typeform	164	25%	8
Jasper	1510	67%	7
Hootsuite	158	53%	8
Salesforce	156	40%	8
Intercom	151	7%	7
CACopy.ai	150	0%	8
SEMrush	150	0%	5
Zendesk	1410	71%	7
Asana	147	50%	7
Mixpanel	144	29%	8
FFreshdesk	143	21%	6
Instapage	140	0%	7
Marketo	135	38%	5
GFGoogle Forms	130	0%	7
Amplitude	126	50%	8
Writesonic	121	8%	7
ZCZoho CRM	120	0%	6
CConvertKit	120	0%	5
Clearscope	120	0%	6
Outreach	115	45%	8

mentions times named first a red rule marks a true zero, not a short bar

The "engines" column is the one that decides whether a zero is meaningful. A tool with no first positions across 8 independent engines is not being missed by one model's quirk.

What to do with this

First · separate the two metrics
Count mentions and firsts apart
A visibility report that gives you one number has averaged away the only distinction that matters here. Ask for appearances and first positions as separate figures, per engine.
If you are mentioned and never first
More mentions is the wrong lever
47 tools in this corpus demonstrate that adding mentions to an already-mentioned tool does not move it up. The tools at the top of the first-rate ladder got there by being the obvious answer to the question, which is a positioning problem rather than a publishing one.
If you are not mentioned at all
Presence first, and it is worth having
Being in the answer puts you in front of a buyer at the moment they ask a purchase question. Just do not let it be sold to you as a route to the top slot.
Either way
Check the category, not the average
In 5 of 16 categories the most-mentioned and most-first tools are different companies. Your own category may be one of them.

Are you mentioned, or are you first?

The free AI Visibility check reports both, per engine, in your own category, with a snapshot date. It is the distinction this page's table is built on.

Run my free AI Visibility check ›

📊 Want the mention-versus-position method? The query set, the 6-mention floor and why it is set there, and the name filter that keeps prose fragments out of the counts. Delivered instantly.

What this does not show

First is list position, not a verdictOn all 288 recorded answers the #1 is the first tool named in the response. It is the position a reader sees first. It is not a separate quality judgement by the engine, and this page does not treat it as one.

A 6-mention floorRate figures use only the 84 tools mentioned 6 times or more. Below that a zero first-rate is explicable by sample size alone, and including it would inflate the headline.

Correlation, not mechanismMentions and first-rate move together at 0.45. Nothing here establishes which causes which, and the most likely answer is that both follow from market position rather than from each other.

One snapshotCaptured in a single window. A tool with a zero first-rate here may lead answers later, particularly in the 5 categories where the top slot is already changing hands.

Bottom line

Being named by an AI engine and being named first by one are close enough to sound like the same product and far enough apart that 47 of the 84 best-covered tools in this corpus have all of the first and none of the second.

Buy presence knowing it is presence. It is worth something. It is not the top of the list, and the 61% figure is what happens when a whole industry prices the two as if they were.

Data: The AI Recommendation Audit (2026), CC-BY, DOI 10.5281/zenodo.20767878. 288 recorded answers, 8 engines, 16 categories, 1561 tool mentions. Related: whether your category is already decided, which engine names the most tools, and which engines cite sources at all.