SG Analytics primary research | AI Visibility Research & Benchmarks | Testing conducted October 2026

Which cybersecurity companies appear in AI- generated vendor answers?

SG Analytics asked ChatGPT, Claude, Gemini, Grok and Perplexity the same 50 cybersecurity buyer questions, then coded every company each answer named and every source it cited. This benchmark shows which vendors AI platforms surfaced, how early they appeared, and whose websites they cited.

ChatGPT Claude Gemini Grok Perplexity
1 10 20 30 40 50
Named at least one company (248)
Named no eligible company (2)
Every tested response: 5 platforms × 50 prompts
50 buyer prompts in 10 intent clusters
5 AI platforms, one run each
250 AI responses analyzed
216 distinct companies observed
2,403 brand appearances coded
4,125 cited URLs classified
01 Key findings

Ten things the benchmark observed

All figures: 50 prompts × 5 AI platforms, 250 analyzed responses, October 2026. Prompts and responses in English; geography. Each rate is shown with its denominator.
92.0% Microsoft mention rate

Three companies appeared in more than 80% of responses

Microsoft appeared in 230 of 250 analyzed responses, CrowdStrike in 213 (85.2%) and Palo Alto Networks in 206 (82.4%). Together they accounted for 27.0% of all 2,403 brand appearances.

46.3% of citations were brand-owned

Under half of cited sources were the named brands’ own sites

1,911 of 4,125 cited URLs pointed to the official domain of a company named in the same response. The share ranged from 9.9% on Grok to 81.3% on Perplexity.

21.7% of third-party citations were listicles

“Best-of” rankings were the most cited third-party page type

Ranking and “best-of” listicles made up 481 of 2,214 third-party citations, followed by review, comparison and directory pages (18.1%). On brand-owned sites, product and service pages led (29.0%).

68.1% Microsoft Top-3 rate

Early placement was concentrated

Microsoft was among the first three companies named in 169 of 248 position-observable responses and first in 78 (31.5%). CrowdStrike was first in 49 and Palo Alto Networks in 41.

44.0% of companies appeared once

Most observed companies appeared only rarely

216 distinct companies appeared in the sample; 95 of them appeared in exactly one response. 31 appeared on all five platforms and 113 (52.3%) on only one.

12.4 vs 6.0 companies per answer

Answer breadth differed by platform

Grok named an average of 12.42 companies per analyzed response and Perplexity 5.96. Claude surfaced 144 distinct companies; Perplexity surfaced 48.

25/25 Category Discovery responses

Prompt intent changed who appeared

Microsoft, CrowdStrike, Palo Alto Networks and Fortinet each appeared in all 25 Category Discovery responses. Fortinet appeared in 2 of 25 Threat / Problem Discovery and 0 of 25 Security Capability Fit responses.

48.0 pts CyberArk cross-platform range

Some companies varied sharply by platform

CyberArk appeared in 68.0% of Grok responses (34 of 50) and 20.0% of ChatGPT responses (10 of 50). Microsoft’s range was 18.0 points.

85.1% specific descriptions

Most brand descriptions carried real detail

2,044 of 2,403 brand appearances were coded Specific. The share ranged from 75.5% on Grok to 93.0% on Gemini.

28.4 vs 7.1 cited URLs per answer

Citation volume differed widely

Analyzed responses carried 4,125 cited URL instances from 759 domains. The average per response ranged from 7.14 on Gemini to 28.38 on Claude.

02 What we asked

50 cybersecurity buyer questions across 10 intents

The prompts follow a buyer’s journey, from broad vendor discovery to threats, security capabilities, industries, regulation, integration, comparisons, operations and trust. Each cluster holds five prompts, and each prompt was asked once on every platform.
01

Category Discovery

37 companies named, 11.88 per answer

Enterprise cybersecurity vendorsCybersecurity platformsEnterprise providersGlobal shortlistEnd-to-end security
02

Threat / Problem Discovery

62 companies named, 8.32 per answer

Ransomware protectionAdvanced threatsPhishingInsider threatsCloud workload protection
03

Security Capability Fit

39 companies named, 7.48 per answer

EDRIdentity securityCSPMXDRAutomated threat response
04

Industry Fit

84 companies named, 11.00 per answer

BankingHealthcareRetailManufacturingTechnology companies
05

Enterprise Profile Fit

58 companies named, 10.48 per answer

10,000+ employeesMid-sized enterprisesGlobal companiesDistributed workforceFast-growing companies
06

Geography & Regulatory Fit

93 companies named, 11.24 per answer

US and EuropeRegulated financial institutionsGDPRHIPAAData residency
07

Technology & Integration Fit

42 companies named, 8.20 per answer

Microsoft environmentsAWSMulti-cloudHybrid cloudSIEM and identity integration
08

Comparison & Alternatives

54 companies named, 10.68 per answer

Vendor comparisonEndpoint security replacementZero trustPlatform vs specialist toolsTool consolidation
09

Implementation & Operations

50 companies named, 7.28 per answer

Enterprise-scale deploymentAlert fatigueSOC automationSmall security teamsSecurity operations
10

Trust & Buyer Confidence

49 companies named, 9.76 per answer

Large-enterprise adoptionAnalyst recognitionCustomer evidenceIndependent validationEnterprise track record
Part 1

Brand visibility in AI answers

Which cybersecurity companies the five AI platforms named, how early, how consistently, and for which buyer questions.

03 Brand visibility

Which cybersecurity companies AI platforms named most often

AI share of voice shows each company’s portion of all 2,403 brand appearances. The eight companies with the most appearances held 48.7% between them; the other 208 companies shared the rest.

AI share of voice

Share of all brand appearances across the 250 responses.
9.6%
8.9%
8.6%
4.9%
4.5%
4.4%
4.1%
3.8%
51.3%
Microsoft 9.57% CrowdStrike 8.86% Palo Alto Networks 8.57% SentinelOne 4.87% Zscaler 4.54% Wiz 4.37% CyberArk 4.08% Okta 3.83% 208 other companies 51.31%

The 20 most frequently observed companies

A company counts once per response, however often it is named. Order reflects appearance counts in this sample only, not product quality or market position.

# Company Mention rate
responses naming the company / 250
AI share of voice
of 2,403
Top-3 first appearance
of 248
First position
of 248
Avg. first position Platforms
1 Microsoft
92.0%230/250
9.57% 68.1%169/248 31.5%78/248 2.96 5/5
2 CrowdStrike
85.2%213/250
8.86% 59.7%148/248 19.8%49/248 3.29 5/5
3 Palo Alto Networks
82.4%206/250
8.57% 52.4%130/248 16.5%41/248 3.35 5/5
4 SentinelOne
46.8%117/250
4.87% 14.9%37/248 1.6%4/248 5.03 5/5
5 Zscaler
43.6%109/250
4.54% 7.3%18/248 4.8%12/248 5.98 5/5
6 Wiz
42.0%105/250
4.37% 10.9%27/248 6.0%15/248 4.83 5/5
7 CyberArk
39.2%98/250
4.08% 4.4%11/248 0.8%2/248 8.05 5/5
8 Okta
36.8%92/250
3.83% 5.6%14/248 0.8%2/248 7.71 5/5
9 Fortinet
32.8%82/250
3.41% 5.6%14/248 0.4%1/248 5.37 5/5
10 Netskope
22.0%55/250
2.29% 3.6%9/248 0.4%1/248 7.94 5/5
11 Splunk
20.8%52/250
2.16% 2.4%6/248 0.4%1/248 7.79 5/5
12 Orca Security
17.6%44/250
1.83% 2.8%7/248 0.0%0/248 8.14 5/5
13 Cisco
17.2%43/250
1.79% 1.6%4/248 0.0%0/248 6.72 5/5
14 Sophos
16.8%42/250
1.75% 3.6%9/248 1.6%4/248 4.95 5/5
15 SailPoint
14.8%37/250
1.54% 1.2%3/248 0.4%1/248 9.27 5/5
16 Proofpoint
13.6%34/250
1.41% 3.6%9/248 1.2%3/248 8.21 5/5
17 Cloudflare
12.0%30/250
1.25% 3.2%8/248 0.0%0/248 5.63 5/5
18 Check Point
11.2%28/250
1.17% 1.2%3/248 0.4%1/248 6.54 5/5
19 Google
11.2%28/250
1.17% 2.4%6/248 0.0%0/248 6.46 5/5
20 Rubrik
10.4%26/250
1.08% 0.0%0/248 0.0%0/248 11.65 4/5
SG Analytics Cybersecurity AI Visibility Benchmark, 50 prompts × 5 AI platforms, 250 analyzed responses, October 2026. Ties ordered alphabetically. “Top-3 first appearance” means among the first three eligible companies named in the answer; it is not a ranking by the AI platform.
04 LLM comparison

ChatGPT vs Claude vs Gemini vs Grok vs Perplexity

Every platform received the same 50 prompts, yet each produced a distinct visibility profile. Differences describe this sample and may reflect model version, retrieval settings and answer length; they are not evidence of a platform preference.

Platform profiles

Signature figures for each AI platform across its 50 responses.

ChatGPT

Avg. companies per answer 7.70
Distinct companies named 58
Avg. cited URLs per answer 25.82
Brand-owned citations 77.8%
Most named Microsoft (46 of 50)
Most cited domain microsoft.com (226 citations)
Most cited page type Product / service page (28.4%)

Claude

Avg. companies per answer 11.48
Distinct companies named 144
Avg. cited URLs per answer 28.38
Brand-owned citations 22.7%
Most named Microsoft (50 of 50)
Most cited domain paloaltonetworks.com (64 citations)
Most cited page type Blog / article (24.8%)

Gemini

Avg. companies per answer 10.50
Distinct companies named 141
Avg. cited URLs per answer 7.14
Brand-owned citations 31.7%
Most named Microsoft (45 of 50)
Most cited domain gartner.com (37 citations)
Most cited page type Ranking / “best-of” listicle (31.7%)

Grok

Avg. companies per answer 12.42
Distinct companies named 69
Avg. cited URLs per answer 10.88
Brand-owned citations 9.9%
Most named Microsoft (50 of 50)
Most cited domain gartner.com (47 citations)
Most cited page type Ranking / “best-of” listicle (20.2%)

Perplexity

Avg. companies per answer 5.96
Distinct companies named 48
Avg. cited URLs per answer 10.28
Brand-owned citations 81.3%
Most named Microsoft (41 of 50)
Most cited domain microsoft.com (87 citations)
Most cited page type Product / service page (27.0%)

Brand coding by platform

Description and role shares use each platform’s brand appearances as the denominator. Every platform has 50 analyzed responses.
Platform Analyzed responses Distinct companies Brand appearances Avg. companies per response Specific descriptions Recommended Compared Alternative
ChatGPT 50 58 385 7.70 87.3% 89.4% 6.0% 4.7%
Claude 50 144 574 11.48 92.9% 76.8% 9.9% 13.2%
Gemini 50 141 525 10.50 93.0% 97.3% 1.5% 1.1%
Grok 50 69 621 12.42 75.5% 59.4% 14.5% 26.1%
Perplexity 50 48 298 5.96 92.3% 90.9% 1.0% 8.1%

Platform-level visibility, 12 most observed companies

Company ChatGPT
n=50
Claude
n=50
Gemini
n=50
Grok
n=50
Perplexity
n=50
Range
Microsoft 92.0%46 100.0%50 86.0%43 100.0%50 82.0%41 18.0 pts
CrowdStrike 90.0%45 88.0%44 80.0%40 94.0%47 74.0%37 20.0 pts
Palo Alto Networks 86.0%43 76.0%38 84.0%42 90.0%45 76.0%38 14.0 pts
SentinelOne 28.0%14 62.0%31 60.0%30 56.0%28 28.0%14 34.0 pts
Zscaler 40.0%20 42.0%21 40.0%20 64.0%32 32.0%16 32.0 pts
Wiz 32.0%16 42.0%21 50.0%25 54.0%27 32.0%16 22.0 pts
CyberArk 20.0%10 32.0%16 48.0%24 68.0%34 28.0%14 48.0 pts
Okta 20.0%10 44.0%22 48.0%24 56.0%28 16.0%8 40.0 pts
Fortinet 38.0%19 24.0%12 28.0%14 42.0%21 32.0%16 18.0 pts
Netskope 14.0%7 10.0%5 12.0%6 58.0%29 16.0%8 48.0 pts
Splunk 30.0%15 44.0%22 12.0%6 8.0%4 10.0%5 36.0 pts
Orca Security 6.0%3 18.0%9 12.0%6 40.0%20 12.0%6 34.0 pts

Share of each platform’s analyzed responses naming the company (count below). Range = highest minus lowest platform rate, in percentage points; single run per prompt, as this is cross-platform variation, not run-to-run variance.

Company ChatGPT
n=50
Claude
n=50
Gemini
n=49
Grok
n=50
Perplexity
n=49
Range
Microsoft 76.0%38 56.0%28 55.1%27 76.0%38 75.5%37 20.9 pts
CrowdStrike 72.0%36 60.0%30 42.9%21 56.0%28 65.3%32 29.1 pts
Palo Alto Networks 68.0%34 44.0%22 44.9%22 46.0%23 59.2%29 24.0 pts
SentinelOne 6.0%3 16.0%8 22.4%11 16.0%8 10.2%5 16.4 pts
Zscaler 4.0%2 4.0%2 10.2%5 12.0%6 6.1%3 8.0 pts
Wiz 12.0%6 10.0%5 12.2%6 10.0%5 10.2%5 2.2 pts
CyberArk 2.0%1 2.0%1 6.1%3 10.0%5 2.0%1 8.0 pts
Okta 4.0%2 4.0%2 8.2%4 10.0%5 2.0%1 8.0 pts
Fortinet 6.0%3 2.0%1 10.2%5 4.0%2 6.1%3 8.2 pts
Netskope 2.0%1 0.0%0 4.1%2 10.0%5 2.0%1 10.0 pts
Splunk 4.0%2 4.0%2 0.0%0 2.0%1 2.0%1 4.0 pts
Orca Security 2.0%1 2.0%1 6.1%3 2.0%1 2.0%1 4.1 pts

Share of each platform’s position-observable responses in which the company was among the first three eligible companies named (count below).

05 Brand overlap

How much the AI platforms agreed on which vendors to name

31 of the 216 observed companies appeared on all five platforms, while 113 appeared on only one. Pairwise overlap shows how similar each platform’s vendor set was.

Companies by number of platforms

How many of the 216 companies appeared on five, four, three, two or one platform.

31
14
20
38
113
All 5 platforms Only 1 platform
On all five (31): AWS, Armis, Bitdefender, Check Point, Cisco, Claroty, Cloudflare, CrowdStrike, CyberArk, Dragos, Fortinet, Google, Microsoft, Mimecast, Netskope, Nozomi Networks, Okta, Orca Security, Palo Alto Networks, Ping Identity, Proofpoint, SailPoint, Saviynt, SentinelOne, Snyk, Sophos, Splunk, Trend Micro, Varonis, Wiz, Zscaler.

Of the single-platform companies, 55 were observed only on Gemini and 48 only on Claude.

Pairwise similarity of vendor sets (Jaccard)

ChatGPT Claude Gemini Grok Perplexity
ChatGPT 58companies 0.33851 / 151 0.32749 / 150 0.53044 / 83 0.58239 / 67
Claude 0.33851 / 151 144companies 0.39080 / 205 0.40161 / 152 0.27241 / 151
Gemini 0.32749 / 150 0.39080 / 205 141companies 0.32952 / 158 0.26039 / 150
Grok 0.53044 / 83 0.40161 / 152 0.32952 / 158 69companies 0.44436 / 81
Perplexity 0.58239 / 67 0.27241 / 151 0.26039 / 150 0.44436 / 81 48companies

Shared ÷ combined companies for each pair of platforms. Diagonal: each platform’s distinct companies.

Most frequent co-mentions

Pairs of companies most often named in the same answer. Co-mention reflects observed association, not competitive equivalence.
Company pair Responses naming both
of 250
Jaccard
CrowdStrike + Microsoft
199 79.6% of responses
0.816
CrowdStrike + Palo Alto Networks
192 76.8% of responses
0.846
Microsoft + Palo Alto Networks
190 76.0% of responses
0.772
CrowdStrike + SentinelOne
115 46.0% of responses
0.535
Microsoft + SentinelOne
114 45.6% of responses
0.489
Microsoft + Zscaler
108 43.2% of responses
0.468
CrowdStrike + Zscaler
105 42.0% of responses
0.484
Palo Alto Networks + SentinelOne
105 42.0% of responses
0.482
06 Buyer intent

Which vendors appeared for each type of buyer question

Category Discovery responses named an average of 11.88 companies but only 37 distinct ones in total; Geography & Regulatory Fit prompts surfaced the most distinct companies (93). Cluster names describe the questions asked, not an endorsement of any company.

Intent-level presence, 10 most observed companies

Analyzed responses in each cluster that named the company (n = 25; 5 prompts × 5 platforms).
Prompt cluster Microsoft CrowdStrike Palo Alto Networks SentinelOne Zscaler Wiz CyberArk Okta Fortinet Netskope
Category Discovery5×25 25 25 25 18 22 19 16 18 25 12
Threat / Problem Discovery5×25 25 15 15 9 7 6 4 5 2 3
Security Capability Fit5×25 25 21 20 15 0 5 6 5 0 0
Industry Fit5×25 17 20 24 7 3 6 15 11 12 1
Enterprise Profile Fit5×25 23 22 16 17 18 9 7 15 9 8
Geography & Regulatory Fit5×25 23 21 18 10 7 11 10 11 1 5
Technology & Integration Fit5×25 21 22 24 9 9 21 10 6 3 3
Comparison & Alternatives5×25 25 21 23 13 17 13 14 14 11 13
Implementation & Operations5×25 23 19 17 9 3 7 2 2 3 2
Trust & Buyer Confidence5×25 23 25 24 10 23 8 14 5 16 8

Where platforms disagreed most

8 cluster cells among the top 10 companies showed the maximum spread: named in all 5 responses on one platform and in none on another.

Wiz

Trust & Buyer Confidence
ChatGPT 0/5
Claude 2/5
Gemini 5/5
Grok 1/5
Perplexity 0/5

SentinelOne

Technology & Integration Fit
ChatGPT 0/5
Claude 4/5
Gemini 5/5
Grok 0/5
Perplexity 0/5

SentinelOne

Enterprise Profile Fit
ChatGPT 3/5
Claude 5/5
Gemini 4/5
Grok 5/5
Perplexity 0/5

Okta

Geography & Regulatory Fit
ChatGPT 0/5
Claude 3/5
Gemini 3/5
Grok 5/5
Perplexity 0/5

Netskope

Trust & Buyer Confidence
ChatGPT 3/5
Claude 0/5
Gemini 0/5
Grok 5/5
Perplexity 0/5

Netskope

Category Discovery
ChatGPT 0/5
Claude 1/5
Gemini 1/5
Grok 5/5
Perplexity 5/5

CyberArk

Trust & Buyer Confidence
ChatGPT 0/5
Claude 3/5
Gemini 4/5
Grok 5/5
Perplexity 2/5

CyberArk

Category Discovery
ChatGPT 3/5
Claude 0/5
Gemini 4/5
Grok 5/5
Perplexity 4/5

Cluster × platform explorer

Choose a company to see how many of the 5 responses per cluster on each platform named it. With 5 responses per cell, small differences of one or two responses shift rankers.

Microsoft: responses naming the company, by cluster and platform

Prompt cluster ChatGPT
n=5
Claude
n=5
Gemini
n=5
Grok
n=5
Perplexity
n=5
spread
Category Discovery 5/5 5/5 5/5 5/5 5/5 0 pts
Threat / Problem Discovery 5/5 5/5 5/5 5/5 5/5 0 pts
Security Capability Fit 5/5 5/5 5/5 5/5 5/5 0 pts
Industry Fit 2/5 5/5 3/5 5/5 2/5 60 pts
Enterprise Profile Fit 5/5 5/5 4/5 5/5 4/5 20 pts
Geography & Regulatory Fit 5/5 5/5 4/5 5/5 4/5 20 pts
Technology & Integration Fit 4/5 5/5 3/5 5/5 4/5 40 pts
Comparison & Alternatives 5/5 5/5 5/5 5/5 5/5 0 pts
Implementation & Operations 5/5 5/5 4/5 5/5 4/5 20 pts
Trust & Buyer Confidence 5/5 5/5 5/5 5/5 1/5 80 pts

CrowdStrike: responses naming the company, by cluster and platform

Prompt cluster ChatGPT
n=5
Claude
n=5
Gemini
n=5
Grok
n=5
Perplexity
n=5
spread
Category Discovery 5/5 5/5 5/5 5/5 5/5 0 pts
Threat / Problem Discovery 3/5 3/5 3/5 3/5 3/5 0 pts
Security Capability Fit 5/5 4/5 5/5 4/5 3/5 40 pts
Industry Fit 5/5 5/5 4/5 5/5 3/5 40 pts
Enterprise Profile Fit 5/5 4/5 4/5 5/5 4/5 20 pts
Geography & Regulatory Fit 4/5 5/5 2/5 5/5 5/5 60 pts
Technology & Integration Fit 5/5 5/5 4/5 5/5 3/5 40 pts
Comparison & Alternatives 4/5 4/5 5/5 5/5 3/5 40 pts
Implementation & Operations 4/5 4/5 3/5 5/5 3/5 40 pts
Trust & Buyer Confidence 5/5 5/5 5/5 5/5 5/5 0 pts

Palo Alto Networks: responses naming the company, by cluster and platform

Prompt cluster ChatGPT
n=5
Claude
n=5
Gemini
n=5
Grok
n=5
Perplexity
n=5
spread
Category Discovery 5/5 5/5 5/5 5/5 5/5 0 pts
Threat / Problem Discovery 3/5 3/5 3/5 3/5 3/5 0 pts
Security Capability Fit 4/5 4/5 4/5 4/5 4/5 0 pts
Industry Fit 5/5 4/5 5/5 5/5 5/5 20 pts
Enterprise Profile Fit 3/5 2/5 4/5 5/5 2/5 60 pts
Geography & Regulatory Fit 5/5 2/5 2/5 5/5 4/5 60 pts
Technology & Integration Fit 5/5 5/5 5/5 4/5 5/5 20 pts
Comparison & Alternatives 5/5 4/5 5/5 5/5 4/5 20 pts
Implementation & Operations 3/5 4/5 4/5 4/5 2/5 40 pts
Trust & Buyer Confidence 5/5 5/5 5/5 5/5 4/5 20 pts

SentinelOne: responses naming the company, by cluster and platform

Prompt cluster ChatGPT
n=5
Claude
n=5
Gemini
n=5
Grok
n=5
Perplexity
n=5
spread
Category Discovery 1/5 4/5 5/5 4/5 4/5 80 pts
Threat / Problem Discovery 1/5 2/5 2/5 2/5 2/5 20 pts
Security Capability Fit 3/5 3/5 3/5 3/5 3/5 0 pts
Industry Fit 0/5 4/5 2/5 1/5 0/5 80 pts
Enterprise Profile Fit 3/5 5/5 4/5 5/5 0/5 100 pts
Geography & Regulatory Fit 1/5 4/5 1/5 4/5 0/5 80 pts
Technology & Integration Fit 0/5 4/5 5/5 0/5 0/5 100 pts
Comparison & Alternatives 2/5 2/5 4/5 3/5 2/5 40 pts
Implementation & Operations 0/5 1/5 2/5 3/5 3/5 60 pts
Trust & Buyer Confidence 3/5 2/5 2/5 3/5 0/5 60 pts

Zscaler: responses naming the company, by cluster and platform

Prompt cluster ChatGPT
n=5
Claude
n=5
Gemini
n=5
Grok
n=5
Perplexity
n=5
spread
Category Discovery 4/5 4/5 4/5 5/5 5/5 20 pts
Threat / Problem Discovery 0/5 3/5 2/5 2/5 0/5 60 pts
Security Capability Fit 0/5 0/5 0/5 0/5 0/5 0 pts
Industry Fit 0/5 0/5 0/5 3/5 0/5 60 pts
Enterprise Profile Fit 4/5 3/5 3/5 5/5 3/5 40 pts
Geography & Regulatory Fit 2/5 1/5 1/5 3/5 0/5 60 pts
Technology & Integration Fit 2/5 3/5 1/5 3/5 0/5 60 pts
Comparison & Alternatives 3/5 3/5 4/5 4/5 3/5 20 pts
Implementation & Operations 0/5 0/5 0/5 2/5 1/5 40 pts
Trust & Buyer Confidence 5/5 4/5 5/5 5/5 4/5 20 pts

Wiz: responses naming the company, by cluster and platform

Prompt cluster ChatGPT
n=5
Claude
n=5
Gemini
n=5
Grok
n=5
Perplexity
n=5
spread
Category Discovery 3/5 3/5 3/5 5/5 5/5 40 pts
Threat / Problem Discovery 1/5 1/5 2/5 1/5 1/5 20 pts
Security Capability Fit 1/5 1/5 1/5 1/5 1/5 0 pts
Industry Fit 1/5 1/5 1/5 1/5 2/5 20 pts
Enterprise Profile Fit 1/5 3/5 2/5 3/5 0/5 60 pts
Geography & Regulatory Fit 3/5 2/5 2/5 2/5 2/5 20 pts
Technology & Integration Fit 3/5 5/5 5/5 5/5 3/5 40 pts
Comparison & Alternatives 2/5 2/5 3/5 4/5 2/5 40 pts
Implementation & Operations 1/5 1/5 1/5 4/5 0/5 80 pts
Trust & Buyer Confidence 0/5 2/5 5/5 1/5 0/5 100 pts

CyberArk: responses naming the company, by cluster and platform

Prompt cluster ChatGPT
n=5
Claude
n=5
Gemini
n=5
Grok
n=5
Perplexity
n=5
spread
Category Discovery 3/5 0/5 4/5 5/5 4/5 100 pts
Threat / Problem Discovery 0/5 0/5 2/5 2/5 0/5 40 pts
Security Capability Fit 1/5 1/5 2/5 1/5 1/5 20 pts
Industry Fit 1/5 4/5 3/5 5/5 2/5 80 pts
Enterprise Profile Fit 0/5 3/5 2/5 2/5 0/5 60 pts
Geography & Regulatory Fit 2/5 0/5 2/5 4/5 2/5 80 pts
Technology & Integration Fit 0/5 3/5 2/5 4/5 1/5 80 pts
Comparison & Alternatives 3/5 2/5 3/5 4/5 2/5 40 pts
Implementation & Operations 0/5 0/5 0/5 2/5 0/5 40 pts
Trust & Buyer Confidence 0/5 3/5 4/5 5/5 2/5 100 pts

Okta: responses naming the company, by cluster and platform

Prompt cluster ChatGPT
n=5
Claude
n=5
Gemini
n=5
Grok
n=5
Perplexity
n=5
spread
Category Discovery 2/5 3/5 4/5 5/5 4/5 60 pts
Threat / Problem Discovery 0/5 2/5 1/5 2/5 0/5 40 pts
Security Capability Fit 1/5 1/5 1/5 1/5 1/5 0 pts
Industry Fit 1/5 4/5 1/5 4/5 1/5 60 pts
Enterprise Profile Fit 3/5 4/5 4/5 4/5 0/5 80 pts
Geography & Regulatory Fit 0/5 3/5 3/5 5/5 0/5 100 pts
Technology & Integration Fit 0/5 2/5 2/5 2/5 0/5 40 pts
Comparison & Alternatives 3/5 3/5 3/5 3/5 2/5 20 pts
Implementation & Operations 0/5 0/5 1/5 1/5 0/5 20 pts
Trust & Buyer Confidence 0/5 0/5 4/5 1/5 0/5 80 pts

Fortinet: responses naming the company, by cluster and platform

Prompt cluster ChatGPT
n=5
Claude
n=5
Gemini
n=5
Grok
n=5
Perplexity
n=5
spread
Category Discovery 5/5 5/5 5/5 5/5 5/5 0 pts
Threat / Problem Discovery 0/5 1/5 0/5 1/5 0/5 20 pts
Security Capability Fit 0/5 0/5 0/5 0/5 0/5 0 pts 0 pts
Industry Fit 3/5 1/5 3/5 2/5 3/5 40 pts
Enterprise Profile Fit 2/5 1/5 1/5 4/5 1/5 60 pts
Geography & Regulatory Fit 1/5 0/5 0/5 0/5 0/5 20 pts
Technology & Integration Fit 1/5 1/5 0/5 1/5 0/5 20 pts
Comparison & Alternatives 2/5 0/5 3/5 3/5 3/5 60 pts
Implementation & Operations 0/5 0/5 0/5 3/5 0/5 60 pts
Trust & Buyer Confidence 5/5 3/5 2/5 2/5 4/5 60 pts

Netskope: responses naming the company, by cluster and platform

Prompt cluster ChatGPT
n=5
Claude
n=5
Gemini
n=5
Grok
n=5
Perplexity
n=5
spread
Category Discovery 0/5 1/5 1/5 5/5 5/5 100 pts 100 pts
Threat / Problem Discovery 0/5 1/5 0/5 2/5 0/5 40 pts 100 pts100 pts
Security Capability Fit 0/5 0/5 0/5 0/5 0/5 0 pts
Industry Fit 0/5 0/5 0/5 1/5 0/5 20 pts
Enterprise Profile Fit 2/5 1/5 1/5 4/5 0/5 80 pts 80 pts 80 pts
Geography & Regulatory Fit 0/5 0/5 2/5 3/5 0/5 60 pts
Technology & Integration Fit 0/5 0/5 0/5 3/5 0/5 60 pts
Comparison & Alternatives 2/5 2/5 2/5 4/5 3/5 40 pts
Implementation & Operations 0/5 0/5 0/5 2/5 0/5 40 pts
Trust & Buyer Confidence 3/5 0/5 0/5 5/5 0/5 100 pts

Part 2

Citations and sources

Whose websites AI platforms cited when answering cybersecurity buyer questions, and what kinds of pages those were.

07 AI citation analysis

Whose websites did AI platforms cite?

Every URL an AI platform supplied in an analyzed response was captured and deduped within that response, giving 4,125 citation instances.

A citation is brand-owned when its domain is the official website of a company named in that same response. Every other link, including another vendor’s site when that vendor was not named, is third-party.

46.3% brand-owned citations
Brand-owned: 1,911 (46.3%) Third-party: 2,214 (53.7%)
ChatGPT 1,291 citations
77.8%
22.2%
Claude 1,419 citations
22.7%
77.3%
Gemini 357 citations
31.7%
68.3%
Grok 544 citations
9.9%
90.1%
Perplexity 514 citations
81.3%
18.7%

Citation distribution by platform

Volume, breadth and ownership of cited URLs per platform. The page type mix by platform appears in section 08.
Platform Cited URL instances Avg. per response Distinct domains Brand-owned Third-party Responses with no citations
ChatGPT 1,291 25.82 99 1,00477.8% 28722.2% 0
Claude 1,419 28.38 494 32222.7% 1,09777.3% 1
Gemini 357 7.14 202 11331.7% 24468.3% 1
Grok 544 10.88 69 549.9% 49090.1% 2
Perplexity 514 10.28 88 41881.3% 9618.7% 2

Brand-owned citations, 12 most observed companies

Own-domain rate = citations to the company’s official domain(s) ÷ all citations in the responses that named it. Share of all brand-owned = its own-domain citations ÷ 1,911.
Company Responses naming it Citations in those responses Own-domain citations Own-domain rate Responses citing its own site Share of all brand-owned citations
Microsoft 230 3,897 345 8.9% 93of 230 18.1%
CrowdStrike 213 3,645 255 7.0% 117of 213 13.3%
Palo Alto Networks 206 3,439 284 8.3% 89of 206 14.9%
SentinelOne 117 1,917 85 4.4% 46of 117 4.4%
Zscaler 109 1,841 90 4.9% 34of 109 4.7%
Wiz 105 1,732 72 4.2% 35of 105 3.8%
CyberArk 98 1,525 18 1.2% 12of 98 0.9%
Okta 92 1,510 33 2.2% 12of 92 1.7%
Fortinet 82 1,385 96 6.9% 33of 82 5.0%
Netskope 55 836 11 1.3% 7of 55 0.6%
Splunk 52 1,158 43 3.7% 16of 52 2.3%
Orca Security 44 669 10 1.5% 7of 44 0.5%

Most cited domains

Citation frequency does not indicate authority or accuracy. Third-party domains exclude the official sites of companies named in the same response.
All cited domains (top 15 of 759)
# domain citations distinct pages platforms
1microsoft.com3452504/5
2crowdstrike.com2541425/5
3paloaltonetworks.com2461655/5
4gartner.com192715/5
5fortinet.com98734/5
6sentinelone.com93664/5
7zscaler.com91724/5
8wiz.io72424/5
9google.com71535/5
10tech-insider.org62233/5
11nist.gov55463/5
12cipherssecurity.com52252/5
13amazon.com49443/5
14splunk.com49373/5
15decryptiondigest.com46123/5
Third-party domains (top 10)
# domain third-party citations
1gartner.com192
2tech-insider.org62
3nist.gov55
4alphanocsecurity.com52
5decryptiondigest.com46
6cybersecuritynews.com42
7esecurityplanet.com40
8thebarrelorgan.com39
9cisa.gov33
10guptadeepak.com28
08 Source type distribution

What kinds of pages AI platforms cited

Each cited URL was assigned one page type (home, product or service, industry blog, listicle, review, news, documentation, research and more), separately for brand-owned and third-party citations.

Brand-owned citations

1,911 citations to named companies’ own sites

Product / service page
29.0%555
Documentation / support
24.3%464
Research, reports & analyst content
10.6%203
Blog / article
8.8%168
Educational / glossary page
6.2%119
Company / investor page
4.1%79
Customer story / case study
4.0%76
Industry, compliance & use-case page
3.2%62
Ranking / “best-of” listicle
3.1%60
Review, comparison & directory page
3.0%57
News / editorial
2.3%44
Home page
1.1%21
Community, social & video
0.2%3

Third-party citations

2,214 citations to all other sites

Ranking / “best-of” listicle
21.7%481
Review, comparison & directory page
18.1%401
Blog / article
17.6%389
Research, reports & analyst content
12.0%266
News / editorial
11.2%249
Government, standards & academic
8.6%190
Documentation / support
3.5%77
Educational / glossary page
1.6%35
Product / service page
1.3%28
Community, social & video
1.1%25
Other
0.9%20
Industry, compliance & use-case page
0.8%18
Company / investor page
0.8%17
Customer story / case study
0.2%5
Home page
0.0%0

Page-type mix by platform

Share of each platform’s cited URL instances by page type (count below). Columns sum to 100% (except for rounding).
page type ChatGPT Claude Gemini Grok Perplexity All
Product / service page 29.4%379 3.2%46 3.4%12 0.0%0 27.6%142 14.0%579
Documentation / support 27.7%358 2.1%30 2.0%7 11.2%61 20.6%106 13.6%562
Ranking / “best-of” listicle 1.7%22 17.9%254 31.7%113 25.2%137 4.5%23 13.3%549
Blog / article 2.3%30 24.6%349 16.5%59 3.9%21 7.8%40 12.1%499
Research, reports & analyst content 10.1%130 10.1%143 16.2%58 18.9%103 6.4%33 11.3%467
Review, comparison & directory page 1.2%15 18.4%261 9.8%35 23.2%126 4.3%22 11.1%459
News / editorial 1.7%22 12.0%170 3.6%13 13.4%73 2.7%14 7.1%292
Government, standards & academic 10.7%138 0.9%13 2.2%8 0.0%0 5.8%30 4.6%189
Educational / glossary page 1.7%22 4.0%57 8.4%30 0.0%0 5.1%26 3.3%135
Company / investor page 2.2%28 3.3%47 0.8%3 3.5%19 3.5%18 2.8%115
Industry, compliance & use case page 4.5%58 0.8%12 2.2%8 0.0%0 4.3%22 2.4%100
Customer story / case study 5.7%74 0.6%8 0.6%2 0.4%2 2.7%14 2.4%100
Community, social & video 0.2%2 1.4%20 0.8%3 0.4%2 0.6%3 0.7%30
Home page 0.7%9 0.0%0 0.3%1 0.0%0 3.7%19 0.7%29
Other 0.3%4 0.6%9 1.4%5 0.0%0 0.4%2 0.5%20
Page types were assigned by documented rules applied to each URL’s path and domain; pages were not opened. In a blind check of 50 randomly selected URLs, the assigned type matched a manual review for 78% (39/50). Most disagreements were between adjacent types, such as a press release filed under a company page or a blog post under news.
Part 3

What it means

Patterns from the data that matter for anyone measuring cybersecurity brand visibility in AI search and AI assistants.

09 Implications

What this means for cybersecurity marketers

Five patterns in this sample, each tied to a measured figure. They describe where AI platforms drew their answers from, not what causes a brand to appear.
39.8%

AI visibility extends well beyond your own website

Ranking, review and comparison pages made up 882 of 2,214 third-party citations in this sample. Tracking how a brand appears on independent listicles, review platforms and comparison pages gives a fuller picture than tracking owned content alone.

53.3%

Product and documentation pages carried most owned citations

When AI platforms cited a named brand’s own site, 1,019 of 1,911 citations went to product, service or documentation pages: 555 product and service pages and 464 documentation pages.

9.9–81.3%

Each AI platform needs its own measurement

The brand-owned share of citations ranged from 9.9% on Grok to 81.3% on Perplexity, and answer breadth from 5.96 to 12.42 companies per response. A single-platform view can misrepresent overall AI visibility.

37 vs 93

Specific buyer questions surfaced more vendors

Broad Category Discovery prompts surfaced 37 distinct companies; Geography & Regulatory Fit prompts surfaced 93. Visibility for narrower, constraint-led questions can look very different from visibility for generic category questions.

192

Analyst content featured prominently

gartner.com was the most cited third-party domain (192 citations), and research, report and analyst pages accounted for 265 third-party citations (12.0%).

10 Glossary

AI visibility terms used in this benchmark

Definitions for the AI search and generative engine optimization terms used throughout this page.

AI visibility

How often, how early and how specifically a brand appears in answers generated by AI assistants and AI search tools, and which sources those answers cite.

Generative engine optimization (GEO)

The practice of improving how a brand is represented in AI-generated answers, as distinct from traditional search engine optimization (SEO) for ranked links.

LLM optimization (LLMO)

Work aimed at how large language models describe, recommend and cite a brand. Often used interchangeably with GEO.

Answer engine optimization (AEO)

Structuring content so that AI assistants and answer engines can extract a direct, accurate answer from it.

AI share of voice

A brand’s appearances as a share of all brand appearances in a defined set of AI responses.

Mention rate

The share of analyzed AI responses in which a brand appears at least once.

First relevant appearance

The order in which an eligible company is first named in an answer. It is an observed placement, not a ranking produced by the AI platform.

Brand-owned citation

A cited URL on the official website of a company named in the same AI response.

Third-party citation

Any other cited URL, such as analyst, media, review, government or other vendor pages.

Buyer-intent cluster

A group of prompts that share a buying question, such as category discovery, industry fit or regulatory fit.

Part 4

Method and reference

How the benchmark was built, how each metric is defined, and answers to common questions

11 Methodology

How the research was done

Which companies appeared in AI-generated answers to enterprise cybersecurity buyer questions, which sources were cited, and how did this differ by platform and buyer intent?

Scope. 50 standardized prompts across 10 buyer-intent clusters, tested once on each of 5 AI platforms (ChatGPT, Claude, Gemini, Grok, Perplexity): 250 responses, all analyzed. Testing period: October 2026. Geography: (geography). Language: English. Access method: (access method). Model versions: (model versions, if recorded).

Brand coding. Analysts recorded each relevant company named in the answer text in order of first appearance, the first three (Top-3), each company’s role (Recommended, Compared or Alternative) and description quality (Specific or Generic, judged on prompt-relevant detail, not factual accuracy). Companies named only inside source URLs, or mentioned only incidentally, were not counted. 2 responses named no eligible cybersecurity company and remain in all denominators with zero brands.

Citation ownership. 119 companies in the sample were matched to 127 official domains, including country domains such as paloaltonetworks.in. Subdomains count with their parent domain; user-hosted content on platform domains (for example github.io or customer apps on azure.com) does not. A citation is brand-owned only when its domain belongs to a company named in that same response. Under this rule, 183 citations to vendor websites count as third-party because that vendor was not named in the response. Ownership was determined by domain; whether each page supports a specific statement about the company was not separately verified.

Name normalization. Spelling variants of the same company were merged (for example AWS and Amazon Web Services; Orca and Orca Security; Google and Google Cloud). Acquired or affiliated businesses named separately, such as Mandiant, Wiz, VMware or Symantec, were kept as separate entries as coded.

Session context. 9 analyzed responses (1 Claude, 8 Perplexity) explicitly referred to earlier questions in the same session, indicating conversation context was not fully reset. They are retained and disclosed here.

Reference material

The exact prompts tested and the full list of metrics reported.
▶ All 50 tested prompts

Category Discovery

  • 1. Which cybersecurity vendors should enterprises consider today?
  • 2. What are the leading cybersecurity platforms for large companies?
  • 3. Who are the major enterprise cybersecurity providers?
  • 4. What cybersecurity vendors should a global enterprise shortlist?
  • 5. Which companies offer strong end-to-end cybersecurity capabilities?

Threat / Problem Discovery

  • 6. Which cybersecurity vendors are strong against ransomware?
  • 7. What solutions help enterprises handle advanced cyber threats?
  • 8. What should enterprises use to reduce phishing risk?
  • 9. Which tools are good for detecting insider threats?
  • 10. What cybersecurity solutions protect cloud workloads from attacks?

Security Capability Fit

  • 11. Which vendors are strong in endpoint detection and response?
  • 12. What platforms should enterprises consider for identity security?
  • 13. Which tools are good for cloud security posture management?
  • 14. What XDR solutions should large enterprises evaluate?
  • 15. Which cybersecurity tools offer strong automated threat response?

Industry Fit

  • 16. Which cybersecurity vendors are well suited for banks?
  • 17. What security solutions work well for healthcare companies?
  • 18. Which cybersecurity platforms suit large retailers?
  • 19. What security vendors should manufacturers consider?
  • 20. Which cybersecurity solutions work well for technology companies?

Enterprise Profile Fit

  • 21. What cybersecurity platforms scale for 10,000+ employee companies?
  • 22. Which vendors are suitable for mid-sized enterprises?
  • 23. What cybersecurity solutions suit global companies?
  • 24. Which security tools work well for distributed workforces?
  • 25. What should a fast-growing company consider for cybersecurity?

Geography & Regulatory Fit

  • 26. Which security platforms work well across the US and Europe?
  • 27. What cybersecurity vendors suit regulated financial institutions?
  • 28. Which cybersecurity solutions help with GDPR requirements?
  • 29. What security platforms should healthcare companies consider for HIPAA?
  • 30. Which vendors support strict data residency requirements?

Technology & Integration Fit

  • 31. Which cybersecurity platforms work well in Microsoft environments?
  • 32. What security solutions work well with AWS environments?
  • 33. Which security platforms are suitable for multi-cloud environments?
  • 34. What should enterprises use for hybrid cloud security?
  • 35. Which cybersecurity tools integrate well with SIEM and identity systems?

Comparison & Alternatives

  • 36. Which cybersecurity vendors should enterprises compare before buying?
  • 37. What options should enterprises consider when replacing endpoint security?
  • 38. Which vendors should companies evaluate for zero trust security?
  • 39. Should enterprises use an integrated security platform or specialist tools?
  • 40. Which vendors help consolidate multiple security tools?

Implementation & Operations

  • 41. Which cybersecurity platforms are easier to deploy at enterprise scale?
  • 42. What tools can help reduce security alert fatigue?
  • 43. Which platforms offer strong SOC automation?
  • 44. What cybersecurity solutions work for small security teams?
  • 45. Which platforms can simplify enterprise security operations?

Trust & Buyer Confidence

  • 46. Which cybersecurity vendors are trusted by large enterprises?
  • 47. Which security platforms are frequently recognized by analysts?
  • 48. Which cybersecurity vendors have strong enterprise customer evidence?
  • 49. What security platforms have strong independent validation?
  • 50. Which cybersecurity companies have a strong enterprise track record?
▶ Metric coverage: 27 metrics reported
Metric Status Where / basis
Total unique brands Calculated 216 companies
Total brand appearances Calculated 2,403
Average brands per response Calculated 9.61
Brand distribution by LLM Calculated Platform comparison
Brand distribution by cluster Calculated Buyer intent section
Most frequently appearing brands Calculated Top 20 table
Brand mention rate Calculated Denominator: 250 analyzed responses
AI share of voice Calculated Denominator: 2,403 appearances
Top-3 first appearance rate Calculated Position coverage 248/250
First position rate Calculated Top 20 table
Average first appearance position Calculated Top 20 table
Intent-level presence Calculated Cluster heatmap
Descriptor quality / specificity Calculated Specific / Generic coding
Co-mention frequency Calculated Overlap section
Total citations Calculated 4,125 URL instances
Unique cited domains Calculated 759 domains
Most cited domains Calculated Citation section
Brand-owned citation rate Calculated 46.3% overall; domain-based
Third-party citation share Calculated 53.7% overall
Source type distribution Calculated URL-based page-type classification
Citation distribution by LLM Calculated Volume, domains, ownership and page-type mix
Cross-platform brand presence Calculated Overlap section
Platform-level mention rate Calculated Platform heatmap
Platform-level Top-3 rate Calculated Platform heatmap toggle
Model variance Calculated Cross-platform range (single run)
Cluster-level platform differences Calculated Cluster × platform explorer
Brand overlap across LLMs Calculated Jaccard matrix
What did SG Analytics measure in this cybersecurity AI visibility study? +
SG Analytics recorded which companies appeared in AI-generated answers to 50 cybersecurity buyer prompts across ChatGPT, Claude, Gemini, Grok and Perplexity in October 2026. Of 250 collected responses, 250 were analyzed. The study measures appearance frequency, first-appearance placement, description specificity, cited sources and source types. It does not assess product quality or security performance.
What is AI visibility, and how is it different from SEO? +
AI visibility measures how often, how early, and how specifically a company is named in synthetic answers by generative models and answer engines, whereas traditional SEO focuses on ranked link positions in search engine result pages (SERPs).
Which cybersecurity companies appeared most often in AI responses? +
Microsoft appeared in 92.0% of analyzed responses, CrowdStrike in 85.2%, and Palo Alto Networks in 82.4%. These top three brands accounted for 27.0% of all 2,403 recorded brand appearances.
How often did AI platforms cite brands’ own websites? +
Across all 4,125 citation URLs classified, 46.3% (1,911 citations) pointed to an official domain owned by a company named in that same response, varying widely by platform from 9.9% on Grok to 81.3% on Perplexity.
What types of pages did AI platforms cite? +
On brand-owned websites, product/service pages (29.0%) and documentation/support (24.3%) led. For third-party domains, “best-of” listicles (21.7%) and review/comparison directories (18.1%) were the most cited.
Do different AI platforms name different cybersecurity vendors? +
Yes. Only 31 of 216 total companies were named across all five platforms, while 113 appeared on only one platform. Breadth also varied significantly, from an average of 5.96 companies on Perplexity to 12.42 on Grok.
Which AI platform cited the most sources? +
Claude cited the highest volume of sources, averaging 28.38 URLs per response across 484 domains, while Gemini averaged 7.14 URLs per response across 202 domains.
How were the 50 prompts chosen? +
The 50 prompts represent a full enterprise cybersecurity buyer journey mapped into 10 intent clusters (5 prompts each), ranging from Category Discovery and Specific Threat Scenarios to Architecture Integration and Trust/Buyer Confidence.
How can cybersecurity brands use these findings? +
Marketing and product teams can benchmark presence against peers, identify which third-party listicles and analyst portals supply engine citations, and structure their documentation and capability pages for higher model retrieval.
What is AI share of voice? +
AI share of voice measures an individual company’s appearances as a proportion of total brand mentions observed across the entire benchmark dataset (2,403 appearances).
Does appearing in AI answers mean a vendor is better? +
No. The study is purely observational and measures algorithmic visibility and retrieval frequency; it does not evaluate software capabilities, security posture, or product efficacy.

See how your brand shows up in AI answers

SG Analytics runs this documented methodology for individual brands and categories: prompt design, multi-platform testing across ChatGPT, Claude, Gemini, Grok and Perplexity, brand coding and citation analysis.

About this research

This benchmark is part of SG Analytics’ AI Visibility Research & Benchmarks series, which measures how brands appear in answers generated by AI assistants and AI search tools. Each study follows a documented publishing standard: standardized prompts, preserved raw responses, reproducible coding and disclosed limitations.

Explore more AI visibility research

How to cite this research

SG Analytics. (2026). Cybersecurity AI Visibility Benchmark: Which cybersecurity companies appear in AI-generated vendor answers? Testing conducted October 2026. https://www.sganalytics.com/ai-visibility/cybersecurity-ai-visibility-benchmark/