The Sales TravelerIndependent · Reader-funded · A partnership buys reach, never a ratingIndex Desk Live
The Sales TravelerThe decision desk for revenue travel
Open the Index Desk
The Index Desk · machine demand

Who's reading this corpus — the machines, counted.

11,931 crawler fetches over 24 day(s). 954 of them (8%) were a person asking an AI a question, and this corpus answering. That number never appears in your analytics.

Window 2026-08-20 → 2026-11-17. User-triggered fetches are the live demand meter: each one is a person asking an AI a question that this corpus answered. Answer engines out-crawl search engines here roughly 2:1. Crawler data only — human traffic is measured separately and stays private.
The split that matters

Three kinds of machine visit.

User-triggered

954

8.0% of crawler fetches. A person asked an AI a question and it fetched a page to answer, in that moment.

Training / ingestion

6,801

57.0% of crawler fetches. The corpus being absorbed into a model or its search index.

Search crawler

4,176

35.0% of crawler fetches. A classic search-engine crawler indexing pages.

User-triggered fetches are the live demand meter. Training crawls measure authority, the corpus being absorbed. Search crawls are the declining channel, kept for contrast.

By engine

Every named engine, in full.

EngineFetch typeFetchesShare
Amazonbot Training / ingestion 3,709 31.1%
Googlebot Search crawler 2,148 18.0%
PerplexityBot Training / ingestion 1,207 10.1%
bingbot Search crawler 824 6.9%
ChatGPT-User User-triggered 824 6.9%
OAI-SearchBot Training / ingestion 746 6.3%
CCBot Training / ingestion 474 4.0%
Applebot Search crawler 409 3.4%
Baiduspider Search crawler 379 3.2%
GoogleOther Search crawler 318 2.7%
GPTBot Training / ingestion 274 2.3%
Meta-ExternalAgent Training / ingestion 166 1.4%
cohere-ai Training / ingestion 132 1.1%
Claude-User User-triggered 86 0.7%
YandexBot Search crawler 55 0.5%
Perplexity-User User-triggered 44 0.4%
DuckDuckBot Search crawler 43 0.4%
Claude-SearchBot Training / ingestion 33 0.3%
ClaudeBot Training / ingestion 29 0.2%
Google-Extended Training / ingestion 23 0.2%
Google-CloudVertexBot Training / ingestion 8 0.1%

Matched against a fixed allowlist. An engine that appears in the data but not the allowlist fails the build rather than being mis-classified.

Coverage

How much of the corpus the machines have reached.

Corpus pages

1,980

Ever crawled

2,537

Coverage

—

Newest first-fetches

Pages an engine reached for the first time, most recent first.

PageFirst fetchedBy
/visiting/kimberly-clark-irving/2026-11-17Googlebot
/sitemaps.xml2026-11-16bingbot
/getting-there/netflix-los-gatos/2026-09-09bingbot
/plan2026-09-09bingbot
/data/airline-delivered.json2026-09-09PerplexityBot
/data/commitment-changes.json2026-09-09PerplexityBot
/data/ad-benchmarks.json2026-09-09PerplexityBot
/data/airport-delivered.json2026-09-09PerplexityBot
/data/conference-drift.json2026-09-09PerplexityBot
/field-guide/devices-data-security/2026-09-09bingbot
/visiting/american-tower-boston/2026-09-09PerplexityBot
/getting-there/moderna-cambridge/2026-09-09PerplexityBot
/visiting/analog-devices-wilmington/2026-09-09PerplexityBot
/getting-there/tjx-framingham/2026-09-09PerplexityBot
/visiting/akamai-cambridge/2026-09-09PerplexityBot
Trend

Machine demand over the window.

DateCrawler fetchesCoverage
2026-08-0223—
2026-08-03198—
2026-08-04306—
2026-08-051,211—
2026-08-061,363—
2026-08-071,717—
2026-08-082,018—
2026-08-092,522—
2026-08-103,096—
2026-08-114,708—
2026-08-125,770—
2026-08-19181—
2026-08-20913—
2026-09-016,585—
2026-09-026,983—
2026-09-037,500—
2026-09-048,782—
2026-09-059,408—
2026-09-069,718—
2026-09-0710,160—
2026-09-0810,725—
2026-11-1711,931—
Where they look

Most-fetched paths.

PathFetches
/robots.txt728
/678
/newsletter-ctas.json164
/intent-concierge.json161
/page-metrics.json157
/insights/where-to-stay-for-dreamforce-2026-best-hotels-in-san-francisco-for-business-travelers51
/sitemap.xml50
/ads.txt36
/changelog/36
/weekly/2026-09-06/33
/sales-ready-index/30
/api/flight-delays30
/flight-delay-report/2026-09-06/30
/insights/flight-delay-tracker/28
/booking-channel-index/28
/flight-delay-report/2026-09-04/25
/insights/25
/business-travel-news/2026-09-04/24
/executive-hotel-index/23
/road-office-index/23
/apartment-stay-index/22
/promised-vs-delivered/jetblue-airways/22
/corrections/20
/commitment-changes/20
/companies/19

Where they fetch from.

CountryFetches
US10,559
CN372
JP372
NL163
KR65
RU46
IN44
DE36
AU31
PL30
ZZ28
CH25
ES25
BE22
SG21
Method & boundaries

How this is measured, and what it never touches.

What is a user-triggered fetch?

When someone asks an AI assistant a question — in ChatGPT, Claude, Perplexity, or a similar tool — and the assistant fetches one of these pages to compose its answer, that request is counted as user-triggered. It is the closest measurable signal to "a person asked an AI about this topic and our work answered." It never appears in a referral log, because the reader usually stays inside the AI tool.

Why is there no human traffic on this page?

This instrument measures machine demand only. Human readership is measured separately under the site’s privacy standard and is not part of this dataset. The split is deliberate and permanent, see the methodology.

Does a crawl change any score or ranking?

No. Demand data — human or machine — may inspire an editorial question, but it never moves a ranking, a finding, a correction priority, or a commercial boundary. A partnership buys reach, never a rating.

Where does this data come from?

First-party, server-side. Every request is counted at the edge and classified by a fixed allowlist of named AI and search crawlers. There is no cookie, no identifier, no stored IP, and no session. The numbers on this page are derived from the committed measurement record and gated at build time.

Independent and reader-funded. The machine-readable instrument is at /data/machine-demand-index.json under a CC-BY licence. Review the independence rules and the verification record. A partnership buys reach, never a rating.