In AI search, see who reads your content — and who sends people back
ChatGPT, Perplexity, Claude and Gemini are changing how people find things. PathClue keeps two ledgers: the visitors AI assistants send you, and the pages their crawlers read.
- AI assistants as their own source
- AI crawlers counted separately
- robots.txt compliance checks
A new traffic source most reports can’t see
Industry reports through 2026 show AI referrals and AI crawler requests growing quickly — while most analytics tools still file them under “referral” and “bots”.
AI answers send visitors
When an assistant cites your page, some readers click through. Those visits convert differently from search — you should know how.
AI crawlers read a lot
Training and answer-engine crawlers fetch thousands of pages. Some send people back; some only take.
You need both numbers
Crawls without visits, or visits without knowing what was crawled, tell half the story. Together they show what your content is worth in AI search.
Two ledgers, read together
Visitors from AI on one side, crawls by AI companies on the other — matched by company, so the ratio is plain.
Crawled vs. sent back
For each AI company: how many pages its crawlers read, how many visitors its assistant sent, and the ratio between them. Spot the crawlers that only take.
- Crawl-to-visit ratio per AI company
- Pages AI sends people to most
- Conversion: AI visitors vs search visitors
| AI company | Pages crawled | Visitors sent | Crawl : visit |
|---|---|---|---|
| OpenAI | 18,420 | 6,830 | 2.7 : 1 |
| Anthropic | 9,310 | 1,120 | 8.3 : 1 |
| Perplexity | 4,120 | 2,390 | 1.7 : 1 |
| 3,880 | 980 | 4.0 : 1 | |
| ByteDance | 2,650 | 0 | ∞ |
AI visitors as a channel
ChatGPT, Perplexity, Claude, Gemini and Copilot each appear as their own source, with landing pages, conversion and revenue — right next to search and social.
- Visitors and conversion per assistant
- Landing pages people reach from AI answers
- Trends you can alert on
| Landing page | Visitors | Conversion |
|---|---|---|
| /pricing | 9,120 | 6.1% |
| /case-studies | 4,310 | 4.8% |
| /blog/… | 12,930 | 1.9% |
Crawl health and compliance
Crawl volume, success rate and errors for every search engine and AI crawler, plus a check of real crawls against your robots.txt and llms.txt — with evidence, block snippets and a rule-file generator.
- Crawl volume, 404s and 5xx per crawler
- Violations of Disallow and Crawl-delay rules
- Generate robots.txt and llms.txt rules
| Crawler | Requests | robots.txt | Status |
|---|---|---|---|
| Googlebot | 6,410 | Follows rules | OK |
| GPTBot | 3,120 | Follows rules | OK |
| Bytespider | 2,650 | Ignores Disallow | Flagged |
| CCBot | 890 | Follows rules | OK |
User-agent: Bytespider Disallow: /
Both ledgers in three steps
AI visitors appear with the tracking script alone. Crawlers rarely run JavaScript, so the crawler ledger needs one server-side source.
- 1
Add the snippet
Visitors from AI assistants show up as their own source from the first visit.
- 2
Connect a server-side source
Send server events through the API or import Nginx / Apache access logs, so crawlers that skip JavaScript are counted too.
- 3
Add your sitemap
PathClue compares your sitemap and robots.txt with what each crawler actually fetched.
AI visibility questions
Anything else — open the chat in the corner and ask.