Traffic you can act on, including the traffic that isn't human.

Status: an idea on the roadmap. Not committed, not scheduled, and no dates.

Page views, sessions, conversion funnels, referrers, attribution — the ordinary questions, answered without a cookie banner. Privacy-friendly by construction rather than by configuration.

The part that is genuinely new

A growing share of the readers of your site are agents, not people. They arrive, read, and decide whether to recommend you — and conventional analytics either discards that traffic as bot noise or never distinguishes it at all.

Agents that authenticate produce unusually legible signal: which agent, which content, how deeply, how often. Knowing which of your pages AI systems actually read, and which they skip, is a question no existing analytics product answers well, and it is becoming a real input into how content gets written.

Why it is separate from Content Workbench

Content Workbench already reports on content — how a piece scores, how it was revised, whether it shipped. That is content analytics.

Web analytics is a different shape of data and a different reader: high-volume event streams, consumed by whoever owns pipeline and revenue rather than whoever owns drafts. Folding the two together would make both worse.

What would have to be true first

Marketing sites generating enough traffic to be worth analysing, and a clear answer on whether cookieless attribution can be made good enough to be honest about. We would rather use an external tool than ship a weak one.

This is a roadmap idea. If you are trying to understand who — or what — is reading your site, we would like to hear how you measure it today.