01 · CRM
HubSpot shows up in 87% of AI answers about CRM. It’s the one to pick in 21%.
87% named
21% the pick
02 · ALL CATEGORIES
0 out of 100: times all five AI engines agreed on which brand to pick
CueScout · 50 buying questions × 2 days on ChatGPT, Gemini, Perplexity, AI Overviews and AI Mode · cuescout.com/research
03 · BY ENGINE
Gemini picked no one in 97% of its answers. Perplexity, in 28%.
Perplexity
28%
ChatGPT
65%
AI Overviews
73%
AI Mode
92%
Gemini
97%
CueScout · Share of answers with no single recommended brand, 100 answers per engine · cuescout.com/research
04 · ONE DAY LATER
Ask ChatGPT the same question a day later and its cited sites overlap by just 23%.
- ChatGPT cited-site overlap
- 23%
- Perplexity cited-site overlap
- 73%
- brand overlap, all engines
- 60%
CueScout · Same 50 questions on October 1 and October 2, 2026 · cuescout.com/research
05 · AUTOMATION
n8n came up in 42 of 100 automation answers. Not one told the reader to pick it.
42%namedpicked: never
CueScout · 100 automation answers from 5 AI engines, October 1–2, 2026 · cuescout.com/research
All five categories
100 answers per category: 10 questions, 5 engines, 2 days. “Named” is the share of answers that mention the brand. “The pick” is the share that tell the reader to choose it.
| Category | Brands named | Named most | Named | The pick | Answers with no pick |
|---|---|---|---|---|---|
| CRM | 72 | HubSpot | 87% | 21% | 61% |
| Sales tools | 101 | Apollo | 70% | 12% | 74% |
| ERP and accounting | 104 | NetSuite | 67% | 9% | 76% |
| Marketing automation | 94 | HubSpot | 65% | 14% | 66% |
| AI agents and automation | 118 | Zapier | 63% | 5% | 78% |
How we did it
50 buying questions, none naming a brand, sent to five engines from a US English desktop location on October 1 and 2, 2026. ChatGPT and Gemini were asked through their consumer apps with web search on, Google AI Overviews and AI Mode came from live Google results, and Perplexity answered through its Sonar API.
Claude Haiku listed the brands each answer named. Jev, a small classifier model, decided which single brand the answer told the reader to pick, or that it picked none. We checked ten of those calls by hand and agreed with all ten. Day-to-day figures are the average overlap between the same engine’s answers to the same question on the two days.
This is a scan made for this report, not customer data. Quote it freely with the sample size attached.