After this read you can decide which buyer questions are still open enough to be worth your next round of work.
Before anything else, we read the questions your buyers actually ask the assistants and check how settled each answer already is. Some questions return the same three brands every time, from sources you will never displace. Others have no settled answer at all. This tells you which questions are worth the work, so you spend your effort where the answer is still open.
Example result
Example. Real measurements, brand anonymised.
Question set read on 12 August 2026. Models: ChatGPT, Gemini and Claude. Unit: share of the answer space still unclaimed, 0 to 100 percent, averaged over 3 runs per model.
What this measures
INPUTS
Up to 20 buyer questions, drawn from your own market and pages.
MODEL SET
ChatGPT, Gemini and Claude, each asked the same question in the same words.
REPEATED RUNS
3 runs per model per question, so a one-off answer cannot set the reading.
OUTPUT
One verdict per question: open room, contested or already settled, with the date of the read.
What you receive
OPEN ROOM
No brand owns this answer yet. Winnable with the pages you already have.
CONTESTED
A shifting set of brands. Position matters more than authority here.
ALREADY SETTLED
The same answer every run, from sources that will not move soon.
Next action: take the open questions into Buyer Match and check whether your pages already answer them.
Limitations
We measure how settled an answer is today. A question that looks open can close when a competitor publishes, so readings are dated and re-run rather than treated as permanent.