What was measured?
Two cybersecurity companies' buyer questions, asked once a day for seven or eight days in October 2026 on seven AI surfaces. Only the questions that name no company are counted, so every vendor in an answer or a search was put there by the assistant. The companies are labelled A (vulnerability management) and B (security awareness training). These are the numbers behind the guide to getting recommended by AI models.
- Named: the vendor's name is in the answer's own words, with links removed.
- Recommendation answer: an answer that names at least three of the category's tracked vendors.
- Own site cited when named: the share of vendor namings where the same answer cites that vendor's own website.
- Top four: the four vendors named most across all seven assistants in each category. The other six per category are combined so the companies stay anonymous.
- Shortlist search: a search asking for the best, top or leading tools, platforms or vendors. Only ChatGPT and Microsoft Copilot expose their searches.
The seven assistants
| AI assistant | Answers | Answers citing a source | Citations per answer | Recommendation answers | Share |
|---|---|---|---|---|---|
| ChatGPT | 1,154 | 99.7% | 18.1 | 201 | 17.4% |
| Microsoft Copilot | 1,162 | 97.4% | 4.6 | 211 | 18.2% |
| Perplexity | 844 | 99.3% | 10.2 | 46 | 5.5% |
| Gemini | 1,162 | 63.1% | 2.1 | 154 | 13.3% |
| Google AI Mode | 937 | 82.6% | 3.7 | 117 | 12.5% |
| Google AI Overviews | 964 | 92.3% | 7.3 | 86 | 8.9% |
| Claude | 1,162 | 7.2% | 0.3 | 432 | 37.2% |
| All seven | 7,385 | 75.8% | 6.5 | 1,247 | 16.9% |
What the assistants searched for
| AI assistant | Shortlist searches | Already named vendors | Share | Added 2026 | Share |
|---|---|---|---|---|---|
| ChatGPT | 93 | 65 | 69.9% | 55 | 59.1% |
| Microsoft Copilot | 138 | 0 | 0.0% | 35 | 25.4% |
Which vendors were named
| Vendor | Category | Named in |
|---|---|---|
| Qualys | Vulnerability management | 738 |
| Tenable | Vulnerability management | 735 |
| Rapid7 | Vulnerability management | 582 |
| ServiceNow | Vulnerability management | 423 |
| Other 6 tracked vendors | Vulnerability management | 118 |
| KnowBe4 | Security awareness training | 1,278 |
| Proofpoint | Security awareness training | 816 |
| Hoxhunt | Security awareness training | 742 |
| Mimecast | Security awareness training | 225 |
| Other 6 tracked vendors | Security awareness training | 546 |
| AI assistant | KnowBe4, security awareness | Tenable, vulnerability management |
|---|---|---|
| ChatGPT | 27.9% | 19.9% |
| Microsoft Copilot | 31.0% | 18.1% |
| Perplexity | 10.5% | 5.8% |
| Gemini | 32.3% | 18.7% |
| Google AI Mode | 32.8% | 21.3% |
| Google AI Overviews | 28.4% | 17.6% |
| Claude | 59.4% | 41.6% |
| AI assistant | All namings | Top four | Other vendors | Other vendor namings |
|---|---|---|---|---|
| ChatGPT | 40.4% | 40.9% | 36.8% | 114 |
| Microsoft Copilot | 9.5% | 4.7% | 32.6% | 172 |
| Perplexity | 37.0% | 34.0% | 54.3% | 35 |
| Gemini | 13.9% | 11.3% | 43.2% | 74 |
| Google AI Mode | 26.3% | 25.3% | 42.5% | 40 |
| Google AI Overviews | 29.2% | 23.7% | 70.1% | 67 |
| Claude | 1.1% | 0.9% | 3.1% | 162 |
| All seven | 16.7% | 14.7% | 32.8% | 664 |
What the assistants cited
| AI assistant | YouTube | Gartner | G2 | PeerSpot | Capterra | TrustRadius | |
|---|---|---|---|---|---|---|---|
| ChatGPT | 1.0% | 0.0% | 20.4% | 34.3% | 4.0% | 3.5% | 4.5% |
| Microsoft Copilot | 0.0% | 0.0% | 0.0% | 7.1% | 0.0% | 0.0% | 0.5% |
| Perplexity | 26.1% | 0.0% | 10.9% | 30.4% | 2.2% | 0.0% | 0.0% |
| Gemini | 5.8% | 0.0% | 42.9% | 9.7% | 0.6% | 0.0% | 0.0% |
| Google AI Mode | 23.1% | 9.4% | 44.4% | 4.3% | 2.6% | 0.0% | 0.0% |
| Google AI Overviews | 51.2% | 38.4% | 30.2% | 7.0% | 3.5% | 0.0% | 1.2% |
| Claude | 0.5% | 0.0% | 0.0% | 1.2% | 0.2% | 0.0% | 0.2% |
| AI assistant | Citations | Roundups and comparisons | Tracked vendors' own sites |
|---|---|---|---|
| ChatGPT | 3,963 | 34.2% | 24.6% |
| Microsoft Copilot | 928 | 66.3% | 10.1% |
| Perplexity | 556 | 52.2% | 22.1% |
| Gemini | 462 | 59.5% | 22.5% |
| Google AI Mode | 598 | 21.6% | 26.6% |
| Google AI Overviews | 634 | 31.5% | 18.1% |
| Claude | 132 | 16.7% | 11.4% |
| All seven | 7,273 | 39.7% | 21.8% |
How the two companies fared
| AI assistant | A named | A's site cited | B named | B's site cited | B named with its own site cited |
|---|---|---|---|---|---|
| ChatGPT | 3 | 99 | 19 | 38 | 14 |
| Microsoft Copilot | 5 | 64 | 1 | 3 | 1 |
| Perplexity | 1 | 23 | 0 | 10 | 0 |
| Gemini | 0 | 7 | 1 | 0 | 0 |
| Google AI Mode | 0 | 4 | 1 | 4 | 1 |
| Google AI Overviews | 0 | 2 | 4 | 4 | 4 |
| Claude | 0 | 0 | 0 | 0 | 0 |
| All seven | 9 | 199 | 26 | 59 | 20 |
What are the limits?
- Two categories, two companies' question sets and about a week of answers.
- Some surfaces returned fewer answers on some days, and Google AI Overviews only answers when Google shows an overview.
- Page types are as the tracking tool classified them, and the searches are the ones the two assistants exposed.