The answer changes before the market does.
Composed answers move faster than rankings and much faster than revenue. That makes them a leading indicator, if you watch the right part of them.
The percentage is a lagging, noisy number. The interesting movement is underneath it, in things that change less often and mean more.
Four leading indicators
- A new name entering the answers at all. A brand appearing for the first time across several questions is a competitor arriving, and it shows up here before it shows up in your pipeline.
- A new source being cited. A page entering the citation list is where the next answers are being decided; getting into it early is far cheaper than getting into it once it is established.
- A question that starts returning consistent names after months of variance. Somebody just claimed it.
- A question that starts hedging after months of confidence. Something destabilised the category’s answer, and the space is briefly open.
None of those requires a statistical argument. They are set changes, countable, and checkable against the stored answers.
Why the percentage is the wrong thing to watch
Because it moves on its own. In a category we measured, a brand at 17.6% weighted share carried an interval from 15.1% to 20.4% across ten samples with nothing changing. Most week-to-week movement in that figure is the instrument, and watching it daily produces a chart that looks like information.
A cadence that works
Monthly readings, with two diffs: which entities are named this month that were not last month, and which sources are cited this month that were not last month. Both lists are usually short, and a short list of real changes beats a long dashboard of movement.
When a new name appears, read the answers it appears in before doing anything. Its description is the reason it was named, and that is more instructive than its size, its funding or its pricing.
The short answers
How can I spot a market opportunity in AI search?
Watch four set changes rather than the score: a new brand entering the answers, a new source entering the citation list, a question that stops varying, and a question that starts hedging. All four are countable and check against stored answers.
Why not just watch the share of voice number?
Because it moves on its own. In a category we measured, ordinary variation ran about five percentage points wide across ten samples with nothing changing, so most short-term movement in it is the instrument rather than the market.
What does a new cited source mean?
That a page has started deciding answers in your category. Getting into it while it is new is materially cheaper than getting into it once it is established.
What does it mean when a question stops varying?
Somebody claimed it. Consistent names on a question that used to return different ones every run is the clearest signal that a competitor has established itself there.
How often should I look?
Monthly, with two diffs: new entities named, and new sources cited. Both lists are short, and a short list of real changes beats a daily chart of noise.
What should I do when a new competitor appears?
Read the answers they appear in first. What the answer says about them is the reason they were named, and it is more useful than their size or their funding.
Find out in under ten minutes
One free check, no account and no card. You get the answers the assistants gave, not a score.
Every figure quoted here comes from a real question put to a real assistant, saved with the date and the assistant that answered it. This piece was written to answer the search "find market gap ai search".