How to read this. AI answers vary by session, personalisation and location, so a single run is a sample rather than a measurement, and we say so rather than dressing it up. Each engine below is marked with how it was checked: queried directly, or verified by hand. Engines marked “pending” were not automated on the date of this run and were not averaged in with the rest.
This is the July 2026 Day-0 baseline for our own studio. Current audits cover ChatGPT, Gemini, Claude and Copilot.
The single most important query, the one a ready-to-hire customer types, and who the AI hands them to.
Every high-intent query across the AI engines we check. Green = named & recommended. Red = the engine names a competitor, not you.
Across every check, this is who the engines named, and how often. Share of voice is the market the AI is currently handing out.
The causes are mechanical. Each one ties to a fixable lever, and those levers are the work that moves you from invisible to named.
Deliberately conservative; real high-intent volume is usually higher. We never overclaim attribution, and AI-search measurement is new enough that we say so. The exact dollar is not the point. Every cell where a competitor is named is work being routed somewhere else.
One-time Foundation installs the plumbing and fixes the causes above. The monthly retainer compounds the signals and tracks every one of these queries so you can watch yourself move.
Entity foundation, schema, GBP overhaul, first answer-pages, and your baseline dashboard. This is the visible win in the first few weeks.
Answer-first content published every month, plus schema, entity and citation upkeep, with a monthly scorecard tracking named-or-not per engine, share of voice, and leading indicators. We report the movement whichever way it goes.
We re-run this same scorecard every month so you can watch the grid go from red to green.