Sales Intelligence · Q2 2026 · Published April 11, 2026

The State of AI Brand Perception in Cybersecurity

This document is not a marketing asset. It is an autopsy.

15Vendors audited 5AI models 75Audits total 33.2sBatch runtime 4Fabrications caught

The cybersecurity industry has a brand problem it did not know it had, because the brand problem is happening inside a distribution layer that its marketing teams do not measure, cannot see, and have never heard of. We measured it. The scoreboard is below. It is not flattering.

The Scoreboard

All scores from a single parallel batch on 2026-04-11, ~17:20 UTC. Five-model council: GPT-4o, Claude Haiku 4.5, Gemini 2.5 Flash, Mistral Large, DeepSeek Reasoner.

#	Domain	Overall	AIPM-NPS	Awareness	Accuracy	Sentiment	Recommend
1	crowdstrike.com	70	+20	85	50	58	85
2	zscaler.com	70	−40	85	50	65	79
3	sentinelone.com	66	−40	68	50	60	79
4	paloaltonetworks.com	56	+25	68	40	60	62
5	darktrace.com	56	−25	68	40	58	68
6	snyk.io	55	0	68	40	58	68
7	rapid7.com	55	0	68	40	58	52
8	recordedfuture.com	54	0	68	40	58	62
9	splunk.com	54	−50	68	40	58	56
10	tenable.com	54	−25	68	40	58	62
11	sophos.com	53	−25	56	40	58	68
12	wiz.io	52	−25	68	40	58	52
13	trellix.com	52	−25	68	40	58	50
14	mandiant.com	51	0	56	40	58	62
15	fortinet.com	49	0	56	40	58	50

Five Observations That Will Ruin a CMO's Weekend

1. Palo Alto has the highest NPS (+25) despite scoring only 56 overall. The brand story is working better than the brand content. The models recommend it but can't produce accurate facts.

2. Splunk has the worst NPS in the batch at −50. Cisco paid $28 billion. Two years later the five largest AI models consider it a net detractor.

3. Fortinet is dead last at 49 overall. A $50B market cap company with $5B+ revenue is scoring below a passing grade in its own industry's AI perception scoreboard.

4. Every single vendor scored between 40 and 50 on accuracy. Not one exception. The entire industry is losing the accuracy fight inside the AI distribution layer and nobody has noticed.

5. Sentiment collapsed into a narrow band. Most vendors scored 58. The models cannot actually differentiate between them. Your differentiation is invisible to the models that are increasingly gating how buyers discover you.

Named Fabrications, Quoted Verbatim

Four errors caught in 33.2 seconds of parallel auditing. Each is reproducible, attributable to a named model, and the kind of error a procurement analyst would never second-guess.

OpenAI GPT-4o

Fabrication #1 — GPT-4o says CrowdStrike is in Sunnyvale

"CrowdStrike was founded by George Kurtz, Dmitri Alperovitch, and Gregg Marston in 2011. The company is based in Sunnyvale, California."

Correct: Austin, Texas (since 2022). Four of five models answered correctly. OpenAI — the largest by user count — was three years stale.

Gemini 2.5 Flash

Fabrication #2 — Gemini invents a founder for Rapid7

"Rapid7 was co-founded by Alan Chhabra and Tas Giakouminakis."

Correct: Alan Matthews, Tas Giakouminakis, and Chad Loder. "Alan Chhabra" appears in zero Rapid7 filings. Gemini fabricated a founder for a NASDAQ-listed company.

Gemini 2.5 Flash

Fabrication #3 — Gemini mutates Snyk's Danny Grander into "Danny Gruss"

"Snyk was founded by Guy Podjarny, Danny Gruss, and Assaf Hefetz."

Correct: Danny Grander. Daniel Gruss is a TU Graz academic (Spectre/Meltdown). Gemini pattern-matched "Danny" + "security research" and substituted the wrong person.

DeepSeek Reasoner

Fabrication #4 — DeepSeek confuses Wiz's Roy Reznik with monday.com's Roy Mann

"Wiz.io was founded by Assaf Rappaport, Ami Luttwak, Yinon Costica, and Roy Mann."

Correct: Roy Reznik. Roy Mann co-founded monday.com — a completely unrelated company. DeepSeek swapped one Tel Aviv founder for another.

Download the full report (PDF)

10 pages. Complete scoreboard. All four fabrications with receipts. Methodology notes.

Download Report ↓

Run Your Own Audit

Every score in this report was produced by the same AIPM API available at aipmsec.com. The audit is reproducible: POST your domain to the endpoint and get the same 5-model council, 4-dimension scoring, and structure signal analysis.

Get an API Key (Free Tier) →

Share This Report

LinkedIn:

https://security.dugganusa.com/reports/ai-brand-perception-cybersecurity-q2-2026?utm_source=linkedin&utm_medium=social&utm_campaign=aipm-q2-2026

Bluesky:

https://security.dugganusa.com/reports/ai-brand-perception-cybersecurity-q2-2026?utm_source=bluesky&utm_medium=social&utm_campaign=aipm-q2-2026

Twitter/X:

https://security.dugganusa.com/reports/ai-brand-perception-cybersecurity-q2-2026?utm_source=twitter&utm_medium=social&utm_campaign=aipm-q2-2026

Email / DM:

https://security.dugganusa.com/reports/ai-brand-perception-cybersecurity-q2-2026?utm_source=email&utm_medium=direct&utm_campaign=aipm-q2-2026

Direct PDF:

https://security.dugganusa.com/whitepapers/state-of-ai-brand-perception-cybersecurity-q2-2026.pdf?utm_source=direct&utm_medium=pdf&utm_campaign=aipm-q2-2026

See how your brand scores

Free AIPM audit: 5 models, 4 dimensions, structure signals. 30 seconds.

Audit Your Domain →