We ran 50,000 prompts through five AI platforms. Here's what gets a brand cited.

Comparison pages beat feature pages 3.2×. Community threads punch far above their traffic. And schema markup matters more than anyone admits.

CITATION RATE BY PAGE TYPE

llms.txt, explained: what it is and whether you need one

A practical walkthrough of the emerging standard, with examples from 40 sites already using it.

Conversation Explorer now covers multi-turn threads

Follow-up questions change which brands get recommended. Now you can see exactly where.

How AI Overviews picks its sources, measured across 8,000 queries

Google's answer engine behaves differently from chat assistants. The overlap is smaller than you think.

Building a prompt set that actually reflects your buyers

Bad prompt sets produce comforting dashboards. Here's how to write ones that hurt — usefully.

Agent Analytics adds llms.txt validation

Catch syntax errors and unreachable paths before crawlers do.

The sentiment gap: when AI mentions you but shouldn't

A mention with stale pricing or a wrong feature list costs more than absence. Measuring the difference.

The 500 domains AI cites most, ranked monthly across five platforms. The category's public benchmark — free, no email wall.

The Answer Layer — monthly research digest

One email a month. Data, not takes.