AI visibility, explained out loud

When a customer looks for a company like yours, who answers?

Two voices take ZenRadar apart, no script read aloud. A growing share of searches ends without a click: an AI answer names three companies, and those are the only ones the customer will ever see. On the agenda: how ZenRadar measures your presence in those answers, how the score out of one hundred is built, and why we leave our own bad numbers on display.

Published August 20, 2026 · Length 5 MIN · updated 2026-08-21 · Listen elsewhere

What the episode covers

• A growing share of searches ends without a click: the person asks an AI, reads the answer, and stops there.
• An AI answer names three companies, sometimes four. Those are the only ones the customer sees, and an absence leaves no trace in your analytics.
• ZenRadar puts your query to five engines: ChatGPT, Perplexity, Gemini, Claude, and the AI overviews inside Google.
• The score out of one hundred breaks down into three pillars: fifty points for citations, twenty-five for the site's markup, twenty-five for the shape of its pages.
• When an overview does not trigger, the report says so instead of counting an absence: a question that was never asked is not a negative answer.

Frequently asked questions

How is this different from Semrush or Ahrefs?

Those tools tell you where you rank in Google. ZenRadar measures whether ChatGPT, Perplexity, Gemini and Claude name you when asked about a company like yours, and which of your competitors they name instead.

Do I need to install anything on my site?

No. The scan reads your public page, the way any engine would. No code to add, no access to grant, no cookie dropped on your visitors.

What happens if a measurement fails?

The report says so, and the scan is not recorded as a result. An unreachable engine would produce a zero that looks like poor performance when it is an outage.