Measurement

How we measure

What Omni Impact checks, which AI engines it asks and how often, and what each number on your dashboard means.

Open in AI

Omni Impact asks AI engines the questions you track and records each answer. It then reports whether the answer names your brand, whether it cites your site, and how it presents you.

What we check

Your plan sets which AI engines answer your tracked questions.

PlanEngines
StarterChatGPT, Google AI Overviews, Google AI Mode, Gemini
ProThe Starter engines plus Perplexity
Premium, Agency, and EnterpriseThe Pro engines plus Claude

Two optional add-ons add engines: the open-model pack (DeepSeek, GLM, Qwen, and Kimi), and Grok from Pro up. See pricing for current plans.

Each answer is saved with its engine, the model that answered, and the date. Your own ChatGPT or Google session can show a different answer, because it uses your chat history and location.

How often

Each engine checks your tracked prompts on its own schedule. Your daily prompts get the faster one.

EngineDaily promptsOther tracked prompts
ChatGPTDailyDaily
Google AI OverviewsDailyDaily
Google AI ModeDailyDaily
PerplexityDailyWeekly
GeminiWeeklyMonthly
ClaudeWeeklyMonthly
Open-model pack (DeepSeek, GLM, Qwen, Kimi)MonthlyMonthly
GrokMonthlyNot checked

Your plan sets how many daily prompts you can mark (see Plans and billing). To check a prompt again sooner, use Execute now in My Prompts. Extra runs count toward your plan's limits.

What the numbers mean

NumberWhat it tells you
Overall scoreYour Website and Prompt scores combined.
Website scoreHow ready your own website is for AI engines, based on a crawl of your site.
Prompt scoreHow strongly AI answers to your tracked questions support your brand, through mentions, citations, and sources.
BPIBrand Perception Index: how answers present your brand, such as how prominently and how favorably.
Engagement scoreA summary of the mentions, citations, and sources in answers to your tracked questions, shown in Prompt Metrics.
Mention rateThe share of answers that name your brand.
Citation rateThe share of answers that cite your site or brand.
Share of voiceYour share of all brand mentions in answers to category questions that don't name you, including brands you don't track. A recommendation counts for more than a passing mention.

Scores are indexes from 0 to 100, not percentages of answers. A rate and a score can move differently because they answer different questions.

Branded questions name your brand and Unbranded questions don't. Prompt Metrics can show each group on its own.

Why numbers move

Your numbers can change when:

  • new answers come in for the selected time range;
  • you edit your prompts, competitors, or brand profile;
  • an engine is added or removed, or moves to a newer model;
  • we improve how answers are analyzed.

Compare the same prompts and time range before reading a change as real movement.

Need more detail?

If you need a deeper walkthrough of how we measure, for example for a procurement review, contact us.

Last updated