For AI agents: a documentation index is available at /llms.txt.
Documentation: Measuring AI visibility

1Alexander / Documentation

Measuring AI visibility

Build a dated record of buyer questions and AI answers. Keep mentions, citations, visits, leads and sales separate, with context and attribution limits.

Reviewed 2026-11-20

Define the questions

Start with a defined set of buyer questions about the product, alternatives, pricing and fit. Check ChatGPT, Google AI Overviews, Perplexity and Grok, explaining any omissions. Keep wording comparable across runs and record changed settings.

Keep the evidence

Record the exact question, named assistant, interface, date and settings where available. Keep the raw answers or enough context to reread them, whether the business was mentioned, cited URLs and who appeared instead. Record when no answer appears. A reviewer needs the evidence behind the summary.

Separate the outcomes

Mentions, citations, referral visits, qualified leads and sales are different records. A mention is not a visit or a sale. Compare dated observations and explain attribution limits before connecting visibility changes to business outcomes.

Read with caution

Answers vary by run. There is no stable rank inside an answer, and a single screenshot is not a ranking. Tools can help retain history, but a visibility score cannot stand in for revenue.

Read Measuring AI search and the AI visibility audit guide for the full record and caveats.