Skip to content
MinuttiaConvex Works

Study overview

MINUTTIA X CONVEX WORKS

AI Search Visibility Data Study

An analysis of 24,067 prompts across Google Search + AIO, ChatGPT, Gemini, Claude and Perplexity.

Analysis snapshot: 26 August 2026

Explore the findingsRead the methodology

Every chart on this page can be downloaded as data, as an image, or shared as a direct link.

24,067

Prompts in the study universe

5

AI search surfaces measured

11,525

Alternative prompts measured on Google

6,363

Comparison prompts analysed

TL;DR

Every question in the study, answered

The 12 research questions behind this report and the short answer to each. Jump to any section for the underlying charts and data.

Summary of every research question and its answer
#QuestionAnswerSection
1Does search across all prompt types, excluding Brand Prompts, trigger an AIO?Yes. 97.7% of alternative and 92.9% of category searches returned an AIO.Q1
2At what percentage do Brand Prompts trigger an AIO?21.5% of brand-name searches returned an AIO.Q2
3Do AIO citations across all prompt types, excluding Brand Prompts, have visibility in Google’s top 10?Partly. 38.5% of AIO citation domains also ranked in the same search’s organic top 10.Q3
4How consistent are results across AI search engines for Category Prompts and Alternative Prompts?Consistency is low. 54% of recommended brands on alternative searches were named by one engine only; 8% by all five.Q4
5What is the average number of brands for Category Prompts and Alternative Prompts?Alternative: 5.5 (Google AIO) to 8.9 (ChatGPT). Category: 7.7 (Google AIO) to 10.5 (ChatGPT).Q5
6Are brands mentioned by AI search engines for Category Prompts and Alternative Prompts also present in AIO?Partly. 35%–54% of an engine’s named brands also appear in the same search’s AIO, depending on engine and prompt type.Q6
7Is Google top-10 visibility associated with mentions in AI search engines for Category Prompts and Alternative Prompts?Yes. Same-category candidate brands ranked in the search’s organic top 10 were 6.8×–11.6× as likely to be named, on every engine.Q7
8Is the number of G2 reviews associated with visibility for Category Prompts and Alternative Prompts?Yes. Within-category top-quartile brands by review count were 5.9× (alternative) and 7.6× (category) as likely to appear in AIO. Star rating shows no comparable association (0.9× on alternative).Q8
9Are products mentioned for Alternative Prompts also cited?Rarely. 3%–29% of brands named on alternative searches carry an official-domain citation, depending on engine.Q9
10Do Comparison Prompts cite either brand’s website?Engine-dependent. 46%–97% of comparison answers cite at least one of the two brands’ sites (ChatGPT 97%, Perplexity 90%, Claude 46%).Q10
11What are the most common evaluation attributes for Comparison Prompts?Features (99%), target fit (94%) and platform scope (91%) appear near-universally; pricing appears in 53% of answers.Q11
12What are the most common questions for Category Prompts and Alternative Prompts?By intent: Best & most popular, Definitions, Top-N lists — together 77% of category follow-ups (63% for alternative). No single exact question dominates.Q12

Questions 01–02

Questions 01–02 continued

AIO trigger rate by prompt type

Share of Google searches that returned an AI Overview. Alternative searches ask for substitutes to a named product; category searches ask for products in a software category.

Alternative prompts: 97.7%; Category prompts: 92.9%; Brand prompts: 21.5%

  • Alternative prompts

    97.7%

    11,265 of 11,525 searches returned an AIO

  • Category prompts

    92.9%

    182 of 196 searches returned an AIO

  • Brand prompts

    21.5%

    1,288 of 5,983 searches returned an AIO

Source: https://study.minuttia.com

Question 03

AIO citations vs. the organic top 10

38.5% of AIO citation domains (39,870 of 103,583 prompt–domain pairs) also appeared in the first ten organic results for the same search. The rate is similar for category and alternative prompts.

Category: 41.7% matched; Alternative: 38.4% matched

Category prompts

Also in organic top 10 — 41.7%Not in organic top 10 — 58.3%

655 of 1,572 AIO citation domains

Alternative prompts

Also in organic top 10 — 38.4%Not in organic top 10 — 61.6%

39,215 of 102,011 AIO citation domains

Source: https://study.minuttia.com

Question 04

How many engines name the same brand

Each prompt–brand recommendation is counted once, by how many of the five engines named that brand on the same search. Only searches answered by all five engines are included.

1 engine: category 51.65%, alternative 53.7%; 2 engines: category 17.83%, alternative 18.59%; 3 engines: category 12.02%, alternative 10.98%; 4 engines: category 9.43%, alternative 8.68%; All 5 engines: category 9.07%, alternative 8.05%

Source: https://study.minuttia.com

Directional overlap between engine shortlists

Each cell covers only searches both engines answered. Percentages are shares of the row engine’s prompt–brand recommendations; overlap is presence in both answers, not evidence of copying.

Prompt type

Directional containment matrix for alternative prompts across 5 engines.

Engine (rows) compared with also named by (columns). The matrix is directional and not necessarily symmetric.
EngineGoogle AIOChatGPTGeminiClaudePerplexity
Google AIO—66.2%53.4%48.5%63.8%
ChatGPT40.9%—38.4%38.4%50.3%
Gemini51.4%59.9%—44.4%56.0%
Claude35.1%44.9%33.3%—46.6%
Perplexity42.3%54.0%38.5%42.7%—

Source: https://study.minuttia.com

Question 05

Average brands per answer

Average canonical brand matches per alternative answer, including answers with no recognized brands.

Google AIO: category 7.7, alternative 5.5; ChatGPT: category 10.5, alternative 8.9; Gemini: category 7.9, alternative 5.7; Claude: category 8.0, alternative 7.6; Perplexity: category 9.9, alternative 8.3

Source: https://study.minuttia.com

Question 06

On an alternative search, ChatGPT names about 8.9 brands per answer; 3.6 of them also appear in Google’s AIO for the same search. The engine-by-engine matrix shows the overlap in both directions.

Overlap between engine recommendations and AIO

Average recognized brands per answer on searches where both the engine and AIO answered, split by whether AIO also named the brand. Overlap is presence in both answers, not evidence of copying.

ChatGPT: category 46.6%, alternative 40.9%; Gemini: category 54.4%, alternative 51.4%; Claude: category 42.7%, alternative 35.1%; Perplexity: category 49.0%, alternative 42.3%

Source: https://study.minuttia.com

Question 07

These are descriptive associations between organic visibility and being named by an AI engine, not causal effects.

Google top-10 rank and mention rate

For each search, candidate brands are all dataset brands in the prompt’s category (excluding the searched brand). Bars compare mention rates for candidates ranked vs not ranked in that search’s organic top 10.

Prompt type

Google AIO: in top 10 33.7%, outside 2.9%, ratio 11.6x; ChatGPT: in top 10 31.8%, outside 4.6%, ratio 6.9x; Gemini: in top 10 28.7%, outside 2.7%, ratio 10.5x; Claude: in top 10 30.3%, outside 3.8%, ratio 7.9x; Perplexity: in top 10 43.5%, outside 4.6%, ratio 9.5x

11.6x

Google AIO — as likely to be named when in the top 10

6.9x

ChatGPT — as likely to be named when in the top 10

10.5x

Gemini — as likely to be named when in the top 10

7.9x

Claude — as likely to be named when in the top 10

9.5x

Perplexity — as likely to be named when in the top 10

Source: https://study.minuttia.com

Question 08

Review counts and star ratings are compared within category. These are associations; the study does not test whether reviews cause AI visibility.

AIO mention rate by G2 review-volume quartile

Within-category quartiles by number of G2 reviews, from the bottom 25% to the top 25%.

Bottom 25%: category 2.0%, alternative 1.4%; 25–50%: category 4.2%, alternative 1.9%; 50–75%: category 6.7%, alternative 3.4%; Top 25%: category 15.5%, alternative 8.5%

7.6x

Category prompts — top vs bottom review quartile

5.9x

Alternative prompts — top vs bottom review quartile

Source: https://study.minuttia.com

AIO mention rate by G2 star-rating quartile

Same structure and scale as the review-volume view, using star rating instead of review count.

Bottom 25%: category 5.0%, alternative 2.9%; 25–50%: category 9.8%, alternative 4.8%; 50–75%: category 8.2%, alternative 4.5%; Top 25%: category 4.7%, alternative 2.6%

0.9x

Category prompts — top vs bottom rating quartile

0.9x

Alternative prompts — top vs bottom rating quartile

Source: https://study.minuttia.com

Review volume and star rating together

AIO mention rate for each combination of review-volume half and star-rating half.

Alternative, Lower half volume and Lower half rating: 1.6%; Alternative, Lower half volume and Upper half rating: 1.7%; Alternative, Upper half volume and Lower half rating: 6.2%; Alternative, Upper half volume and Upper half rating: 5.5%; Category, Lower half volume and Lower half rating: 2.5%; Category, Lower half volume and Upper half rating: 3.4%; Category, Upper half volume and Lower half rating: 12.4%; Category, Upper half volume and Upper half rating: 9.7%

Category prompts

Category prompts: AIO mention rate by review-volume half and star-rating half
Review volumeLower half ratingUpper half rating
Lower half2.5%3.4%
Upper half12.4%9.7%

Alternative prompts

Alternative prompts: AIO mention rate by review-volume half and star-rating half
Review volumeLower half ratingUpper half rating
Lower half1.6%1.7%
Upper half6.2%5.5%

Source: https://study.minuttia.com

Mention coverage

How many mentions were eligible for the G2 comparison.

Mention coverage by prompt type
Prompt typeAll mentionsCategory-eligibleAnalysed
Alternative61,66128,66021,081
Category1,405628460

Question 09

Official domain cited by engine

Share of recognized mentions accompanied by an official-domain citation.

Prompt type

Google AIO: 12.2%; ChatGPT: 28.6%; Gemini: 3.2%; Claude: 5.5%; Perplexity: 15.7%

Source: https://study.minuttia.com

Official domain in Google top 10

Share of recognized mentions whose official domain also ranked in Google’s top 10.

Prompt type

Google AIO: 9.7%; ChatGPT: 6.4%; Gemini: 7.3%; Claude: 6.5%; Perplexity: 8.8%

Source: https://study.minuttia.com

Question 10

Do comparison searches cite either brand’s website?

Each brand was paired with its closest competitor and every engine was asked 6,363 “How do X and Y compare?” questions. Engines almost always name both brands (95.8%–98.5% of answers) but rarely cite either brand’s website: 96.8% of ChatGPT answers and 45.6% of Claude answers link to at least one of the two sites.

Answers counted

Google AIO: both 37.3%, subject only 27.1%, competitor only 19.7%, neither 15.9%; ChatGPT: both 77.2%, subject only 9.9%, competitor only 9.7%, neither 3.2%; Gemini: both 13.7%, subject only 27.0%, competitor only 15.1%, neither 44.3%; Claude: both 11.3%, subject only 20.9%, competitor only 13.4%, neither 54.4%; Perplexity: both 49.9%, subject only 23.2%, competitor only 17.4%, neither 9.5%

Source: https://study.minuttia.com

Own-domain citation rates by engine

Share of comparison answers citing the subject brand's domain, the competitor's domain, and either of them.

Google AIO: either 84.1%; ChatGPT: either 96.8%; Gemini: either 55.7%; Claude: either 45.6%; Perplexity: either 90.5%

Source: https://study.minuttia.com

Who do the engines favor?

ChatGPT, Perplexity and Claude recommend one of the two brands in 46%–50.7% of answers; Gemini and Google AIO usually stay balanced (82.4% and 76.9%).

Google AIO: subject 7.7%, competitor 14.0%, balanced 76.9%; ChatGPT: subject 16.0%, competitor 34.7%, balanced 49.2%; Gemini: subject 5.6%, competitor 11.6%, balanced 82.4%; Claude: subject 14.5%, competitor 31.5%, balanced 44.3%; Perplexity: subject 13.3%, competitor 30.7%, balanced 55.4%

Source: https://study.minuttia.com

Domains most often cited in comparison answers

Top 10 cited domains for the selected engine, by share of comparison prompts.

Engine

youtube.com: 62.4%; reddit.com: 44.0%; g2.com: 19.6%; linkedin.com: 11.4%; quora.com: 9.7%; 6sense.com: 9.6%; facebook.com: 9.2%; medium.com: 8.4%; peerspot.com: 7.1%; trustradius.com: 6.8%

Source: https://study.minuttia.com

Google AIO comparison coverage

Collection context for the Google AIO comparison sample.

Google AIO comparison coverage measures
MeasureValue
prompts6,363
serp ok6,363
aio triggered raw6,075
aio with content3,841
aio trigger rate raw95.5%
aio content rate60.4%
aio answers with citations3,837
avg share aio citations in organic top1067.0%
share aio answers overlapping organic99.3%

Question 11

Which evaluation attributes do comparison answers weigh?

Share of answers discussing each attribute. One answer counts toward every attribute it discusses.

View

Features: 99.0%; Target fit: 94.5%; Platform scope: 90.5%; Integrations: 66.2%; Ease of use: 59.0%

Source: https://study.minuttia.com

Which attribute opens the answer?

The first attribute each answer discusses.

Engine

Platform scope: 48.9%; Target fit: 13.3%; Features: 10.7%

Source: https://study.minuttia.com

Attribute emphasis by category

Share of comparison answers in each software category that discussed the attribute.

Category

Features: 99.7%; Target fit: 97.9%; Platform scope: 88.7%

Source: https://study.minuttia.com

Behaviours around the comparison itself

Share of comparison answers that frame the two products as complementary, and share that caveat or cite their evidence.

Google AIO: complementary 8.9%, evidence caveat 5.5%; ChatGPT: complementary 29.9%, evidence caveat 29.4%; Gemini: complementary 15.1%, evidence caveat 1.4%; Claude: complementary 8.9%, evidence caveat 30.3%; Perplexity: complementary 13.4%, evidence caveat 39.3%

Source: https://study.minuttia.com

Question 12

Most common buyer questions

Google’s People Also Ask entries collected alongside category and alternative searches. Individual exact questions repeat infrequently; classified by intent, they concentrate into a few groups.

Prompt type

Best & most popular: 31.6%; Definitions: 23.6%; Top-N lists: 22.0%; Which & who: 10.1%; Yes/no evaluations: 7.4%; How-to & usage: 3.1%; Pricing & cost: 1.7%; Comparisons: 0.5%

Source: https://study.minuttia.com

Most frequently repeated exact questions

No single phrasing dominates: the most repeated exact question appears only a handful of times.

Prompt type

What are the top 10 SOC tools? (8 appearances); What are the top 5 EDR tools? (7 appearances); What are the 5 marketing tools? (6 appearances); What are the top 10 analytics tools? (6 appearances); What are the 5 best tools for performance management? (5 appearances)

Most common category follow-up questions
RankQuestionPrompt typeAppearances
1What are the top 10 SOC tools?Category8
2What are the top 5 EDR tools?Category7
3What are the 5 marketing tools?Category6
4What are the top 10 analytics tools?Category6
5What are the 5 best tools for performance management?Category5
6What are the top 3 CRM tools?Category5
7What are the top 5 automation tools?Category4
8Is there a better AI than ChatGPT?Category3
9What are examples of CX tools?Category3
10What are the 5 main AI tools?Category3
11What are the 5 most popular AI tools?Category3
12What are the big 3 AI tools?Category3

Source: https://study.minuttia.com

Additional analysis

Brand-name searches

What AI surfaces return when the search is the brand's own name.

5,983

Brand-name searches measured

21.5%

Returned an AIO

1,288

Brand searches with an AIO

93.2%

Own domain in Google's organic results

Competitor presence and own-domain citations on brand searches

For each engine: how often a competitor was named in the answer, and how often the brand's own domain was cited.

Google AIO: competitor named 49.5%, own domain cited 70.3%; ChatGPT: competitor named 30.2%, own domain cited 82.9%; Gemini: competitor named 35.8%, own domain cited 50.3%; Claude: competitor named 13.1%, own domain cited 24.8%; Perplexity: competitor named 32.6%, own domain cited 96.3%

Source: https://study.minuttia.com

Brands most often named alongside the searched brand

The most-named brands skew toward large platforms referenced as integrations, app stores, or infrastructure. This measures presence in other brands’ answers, not recommendations to switch.

Engine

YouTube: 304 searches; Reddit: 53 searches; Google: 33 searches; GitHub: 28 searches; Microsoft: 28 searches; Kubernetes: 23 searches; Google Cloud: 20 searches; LinkedIn: 15 searches; Microsoft Teams: 14 searches; QuickBooks: 11 searches

Source: https://study.minuttia.com

Claude's response behaviour

Given only a brand name, Claude asked for clarification instead of answering on 46.4% of prompts. Every other engine searched and answered essentially every brand prompt.

Claude response behaviour on brand-name searches
BehaviourSearchesShare
Asked to clarify without searching2,60043.5%
Searched the web and answered2,41440.4%
Answered from model knowledge79413.3%
Searched, then asked to clarify1752.9%

Additional analysis

Category detail

The same measures broken out by software category. Categories with fewer than 30 alternative prompts were excluded to keep denominators meaningful.

154

Categories included

42

Categories dropped for low volume

30

Minimum alternative prompts per category

Top 15 categories by named by all five engines

Share of the category's recommended brands named by all five engines

Measure

Data Labeling: 18.2%; Enterprise Resource Planning: 15.3%; Point of Sale & Retail: 15.0%; Database Management Systems: 14.7%; 3D Modeling: 13.5%; Data Lake/Warehouse: 13.1%; Privileged Access Management: 12.5%; Contract Lifecycle Management: 12.1%; Video Meetings: 11.8%; Cloud Security Posture Management: 11.8%; Office Productivity Suites: 11.7%; Help Desk: 11.6%; CI/CD Platforms: 11.4%; Version Control: 11.3%; Graphic Design: 11.1%

Source: https://study.minuttia.com

Additional analysis

Most cited domains by engine

The sources AI answers lean on, with each domain's type as classified in the dataset.

Most frequently cited domains

Share of the engine's citing prompts where the domain appeared as a citation.

Engine

g2.com (Review platform): 39.2%; reddit.com (Community): 38.4%; youtube.com (Video): 35.5%; gartner.com (Analyst): 19.2%; linkedin.com (Social): 6.7%; medium.com (Editorial): 6.2%; checkthat.ai (Software vendor): 5.5%; zapier.com (Vendor content): 5.3%; capterra.com (Review platform): 5.1%; learn.g2.com (Review platform): 4.8%

Source: https://study.minuttia.com