How often are you cited?
A living report on citation rates across the seven AI engines that matter for B2B discovery. 200 prompts, 7 engines, refreshed every month.
Curated B2B panel across 6 categories.
ChatGPT, Perplexity, Gemini, Claude, Copilot, Grok, AIO.
Twice-weekly runs across the June measurement window.
Citation behaviour by engine
Across identical prompts, engines differ by an order of magnitude in how many sources they cite. If you optimize for citation share, Perplexity, AIO, and Copilot are the highest-leverage targets today.
| Engine | Avg citations / answer | Answers with ≥1 citation | Unique domains / 100 answers | Wikipedia % | Reddit % | YouTube % |
|---|---|---|---|---|---|---|
| Perplexity | 6.2 | 99% | 214 | 14% | 11% | 4% |
| Google AI Overviews | 4.8 | 97% | 188 | 9% | 16% | 12% |
| Copilot | 4.1 | 94% | 172 | 12% | 9% | 6% |
| ChatGPT (Search) | 3.4 | 82% | 141 | 17% | 13% | 5% |
| Grok | 2.9 | 74% | 118 | 6% | 8% | 3% |
| Gemini | 2.6 | 71% | 122 | 18% | 7% | 9% |
| Claude | 1.1 | 34% | 62 | 22% | 4% | 2% |
Citation & brand-mention rate by query category
Not all queries behave the same. B2B SaaS listicles get named brands 88% of the time; regulatory queries almost never do — engines route to institutional sources.
| Category | Avg citations | Brand-answer rate | Listicle share | Vendor-site share |
|---|---|---|---|---|
| B2B SaaS ("best X software") | 5.4 | 88% | 46% | 21% |
| Dev tools & API comparisons | 4.9 | 84% | 32% | 27% |
| Local services (dental, HVAC, legal) | 3.1 | 61% | 14% | 8% |
| E-commerce product picks | 4.2 | 79% | 41% | 12% |
| How-to & tutorials | 3.7 | 44% | 22% | 6% |
| Regulatory / medical / legal facts | 5.1 | 12% | 8% | 2% |
What this means for GEO strategy
- Prioritize engines that cite. Perplexity, AIO, and Copilot together account for the vast majority of clickable citations. Optimizing for Claude is largely a brand-mention play, not a traffic play.
- Get on Reddit and Wikipedia. They dominate as source domains across every engine except Grok. A single well-cited Reddit thread often outranks a vendor blog for years.
- Listicles still work. Third-party "best X" round-ups capture 32-46% of the citation pool for B2B queries. Earning placement in three strong listicles per category outperforms most on-site SEO tactics in AI surfaces.
- Category matters more than tactics. Regulatory and medical categories will not surface brand names no matter how well you optimize. Do not budget for GEO in those verticals until engines shift policy.
Methodology
The 200-prompt panel spans six B2B-relevant categories: SaaS comparisons, developer tools, local services, e-commerce picks, how-to tutorials, and regulatory/medical facts. Prompts were frozen on May 30 2026 and re-issued twice weekly through June to each of the seven engines via their public interfaces (Perplexity, ChatGPT with search, Gemini, Claude with web tool, Copilot, Grok, and Google AI Overviews). Response HTML was parsed to extract cited URLs, which were normalized to registered domains and de-duplicated per answer. Brand-mention rate is computed against the full brand set defined per category, not just the answer's cited sources.
Report is CC-BY 4.0. Cite as: Optymia, "How often are you cited?", Q3 2026 report, 2026-07-06. The dataset refreshes monthly; the next update posts on 2026-08-06.
Want your brand added to the panel?
Optymia customers on Growth and above get their brand tracked automatically against the full prompt panel with a live citation-share dashboard.