Where does a market's question set come from?

Every market is measured with a fixed question set: 10 to 15 questions that KI Score uses to measure every brand in that market alike. This article explains where those questions come from, how they are selected, and why you can't view them in the dashboard.

Search demand as the foundation

The question set isn't made up from scratch, apart from one exception when demand is very low (see below). It's based on what people in a market actually search for. The starting point is a curated list of search terms for each market plus demand data, including data from SISTRIX, which KI Score uses for demand and topic research. Google Search suggestions are added as well, as long as they're phrased as questions. The questions are written in the market's language.

This demand data is used only to build the question set. KI Score doesn't show search volumes or real conversations between users and AI systems.

How search queries become a question set

From the search queries it finds, KI Score picks the ones that work for a fair market comparison:

• Recency: The questions should reflect current demand.
• Relevance and market fit: Every question has to match the market's service and geographic market.
• No brand names: Questions that already name a brand are dropped, so no brand gets a head start from the way a question is phrased.
• Useful for comparing brands: An answer to the question should be likely to name several brands or providers. Pure definition, price, or personal advice questions are dropped.
• No duplicates: Of several questions that mean the same thing, only one is kept, even if they're worded differently.

Once a question set is activated, it stays unchanged. Any change creates a new version of the set and therefore a new measurement series. That way, the KI Score always compares measurements that use the same questions within a series.

10 to 15 questions with different intents

A question set contains 10 to 15 questions covering different question intents. The target mix is:

The actual mix can differ when a market doesn't have enough matching demand for an intent. To learn what each intent means and how it feeds into the KI Score, see the article on question intents.

Why the question set stays private

Only the KI Score team can see the individual questions and the underlying search volumes, internally. No customer role has access, not even workspace admins. The KI Score page shows neither the individual questions nor their mix of intents nor the raw answers from the market measurement. Even the progress of a running measurement appears only as a share of answers, not as a count.

The question set reflects demand in the market and is the same for every brand. It's the shared basis of the market comparison, not part of your own settings.

If you want to track particular questions yourself, add them as your own prompts in monitoring. Your own prompts don't change the KI Score, but they show you how AI systems answer exactly those questions.

When there isn't enough demand

If a market yields fewer than 10 usable questions, it's considered not fairly measurable. KI Score then measures one geographic level up, for example the region instead of the city, or the country instead of the region. Your selection in the market profile stays as it is.

If demand still isn't sufficient at the larger levels, KI Score adds questions written by an AI without demand evidence to the question set for the level you chose, up to 12 questions in total. The questions from real demand stay first. Only if that doesn't work either does the KI Score page show the notice Zu wenig messbare KI-Nachfrage (Too little measurable AI demand). Your monitoring with your own prompts keeps running regardless.