Measure

Engines and markets

Which AI engines Rixe measures, how your market decides which ones apply to you, and how to measure a brand in more than one country.

Rixe measures eleven AI answer engines. Which of them are measured for you depends on two things: your plan sets how many, and your market — a country and a language — decides which ones, according to how much each engine is used there. This page lists the engines, explains what a market pack contains, and shows how to open a second market.

The engines

EngineSearches the webCites sourcesCounts toward the visibility score
ChatGPT (Search)✓✓✓
Gemini (Grounding)✓✓✓
Perplexity (Sonar)✓✓✓
Google AI Overviews✓✓✓
Google AI Mode✓✓✓
Microsoft Copilot✓✓✓
Grok (xAI)✓✓✓
Claude (Sonnet)———
DeepSeek———
Mistral (le Chat)———
Qwen———

Search-grounded engines run a live web search before answering and report the pages they used. They fill the Sources view and are the engines that count toward the composite visibility score.

Chat models (Claude, DeepSeek, Mistral, Qwen) answer from training data without searching. Their answers are measured and shown per engine — useful to see what a model "believes" about your category — but they never cite sources and carry no weight in the composite score, because a model without search rarely names a niche or local brand however well that brand ranks.

A few engine-specific notes:

  • Google AI Overviews does not appear for every search. When Google shows no AI answer for a prompt, that is recorded as a run in which no brand was named — a measurement result, not a failure.
  • Microsoft Copilot is measured through the Bing answer surface, because Copilot has no public query interface.

Which engines are measured for you

You do not pick engines by hand. Rixe ranks the engines by their weight in your market pack and measures the top ones, up to your plan's number:

PlanEngines measured
Solo, Growth, ScaleThe 3 highest-weighted engines in your market
Leader and all agency plansEvery engine your market includes

The setup wizard's Engines step shows the exact list for your market before you finish. The same engines are measured every day, in the same order, so your results stay comparable over time.

If an engine in your set cannot measure on a given night, the screen says so: setup leaves it out of the first scan with a note, a banner on AEO Insights › Visibility names degraded engines, and their missing answers are left out of your rates rather than counted as misses. See How measurement works.

Markets

A market is a country together with a language — Germany · German, Canada · French, United Arab Emirates · Arabic. Language matters as much as country: people in Belgium ask in Dutch or French, and the answers differ. Each market is measured on its own questions, in its own language, on the engines that matter there.

Market packs

Everything Rixe knows about a market lives in its market pack, not in code. Rixe ships 98 market packs across 48 languages. A pack includes:

  • Engine weights — how much each engine is used in that market. They decide which engines a plan measures and how much each counts in the visibility score. An engine can be deliberately excluded from a market.
  • Language rules — how brand names are recognised in answers: letter case, accents, apostrophes and suffixes. In Turkish, for example, "Trendyol'da" and "trendyolda" are both recognised as the brand Trendyol.
  • Local sources — the sites that feed AI answers in that country (forums, review sites, marketplaces, news, directories), used by Sources and the action plan.
  • Review and listing platforms for that market, and other local rules such as how page types and dates are written.

Opening another market

Each market is its own workspace, so the number of markets you can open equals your plan's workspace limit (see Plans and limits). You can choose several markets in the setup wizard, or add one later:

  1. Go to AEO Insights › Markets.
  2. Click Add market.
  3. Choose the country and language under Market package.
  4. Click Open this market.

The new workspace gets the same brand and competitors, and the prompts tracked in your current workspace are copied across ("Market opened: 48 prompts and 6 brands carried over"). The copied prompts count toward your plan's prompt limit. Review them in the new workspace: a question worded for one country is not always how people ask in another country or language, and measuring the same words as if they meant the same thing produces a score that means little.

Markets are compared, never averaged

AEO Insights › Markets shows your markets side by side. Rixe does not combine them into one number: a brand can be strong in Türkiye and invisible in Germany, and the average would describe neither.

Switching between markets

Use the workspace switcher, or the market filter on AEO Insights, to look at one market at a time.

Common questions

Can I add an engine my plan does not measure? Brand plans measure the three highest-weighted engines in your market; Leader and the agency plans measure every engine. To measure more engines, move to one of those plans.

Why is an engine I use personally not in my list? Its weight in your market is lower than the engines above it. The weights reflect how much each engine is used in that country.

Why do chat models not affect my score? Without a web search they rarely name brands that are not already famous, so counting them would make a well-ranked brand look invisible. Their answers are still shown per engine.

Still stuck? Write to [email protected] and a person will answer.