Site index for AI agents: /llms.txt. Documentation pages under /docs are also served as markdown at the same URL plus .md (e.g. /docs/geo-audit.md) or via an Accept: text/markdown request.

Your analytics calls it direct. It was ChatGPT.

Answer engines send visitors without a referrer, so the channel hides in direct. AirPulse separates it by engine and page.

Fund the channel AI is already sending.

The pages engines send people to are rarely the pages you optimised.
AI visitors arrive later in the cycle, so raw session counts undersell them.
AirPulse overview showing visibility signals and prioritized prompts

Measure AI traffic from volume to value.

Attribution

Referrals from ChatGPT, Perplexity, Gemini, Copilot and Google AI, pulled out of the direct bucket. Sessions, trend and share of total.

Engine breakdown

Which engines send people, and in what mix. The engine that cites you most is rarely the one that sends the most traffic.

Landing pages

The pages engines send people to, which tells you where to put the demo CTA.

Quality signals

Time on page, pages per session and bounce for AI visitors against the rest of your traffic.

An engine can cite you, answer the buyer's question completely, and send nobody. No traffic report will show it. Read AI Traffic next to visibility, never instead of it.

Questions teams ask first

Most answer engines send visitors without a referrer, or with one analytics does not recognise as a search source, so the visit lands in direct. AirPulse identifies the referral patterns each engine uses and pulls them back out.

Measure, act, grow.

See all features

Turn AI visits into pipeline.

A walkthrough on your own analytics: engines, landing pages and visit quality.