See where your brand stands in AI answers, what changed, and what needs attention now.
Performance is the understanding layer. It turns stored AI answers into reproducible numbers for visibility, recommendations, citations and share of voice, and always shows the sample size, scope and evidence behind them.
Performance
What is happening?
- Overview
- AI Visibility
- AI Answers
- Sources
- Competitors
- Search
- Business Impact
Quick answer: Performance
Performance is the understanding layer of BrandWater AI. It runs your activated prompts on Gemini, Google AI Overview, Google AI Mode, ChatGPT and Perplexity, stores every answer in full, and reports visibility, recommendation rate, citation rate and share of voice with the sample size and evidence shown. Search Console and GA4 views are coming in V2.
Metric explorer
Choose a capability to see the numbers it produces and exactly how each is defined.
How is my brand doing in AI discovery right now?
The default screen. It follows one order every time: current state, what changed, why, the evidence, and what needs attention next.
Open Overview- Current state
- The latest snapshot for the selected engines, market and window.
- Change
- Difference against the prior comparable window, shown only when the sample supports it.
- Attention items
- The few opportunities with the most impact and confidence, never a long list.
What Performance gives you
7 capabilities, each answering a specific question and labelled with its release stage.
How Performance measures things
Every metric has a definition, so a number always means the same thing.
| Metric | Definition | Capability |
|---|---|---|
| Current state | The latest snapshot for the selected engines, market and window. | Overview |
| Change | Difference against the prior comparable window, shown only when the sample supports it. | Overview |
| Attention items | The few opportunities with the most impact and confidence, never a long list. | Overview |
| Mention rate | Runs that mention the brand divided by eligible runs. | AI Visibility |
| Recommendation rate | Runs that recommend the brand divided by eligible runs. | AI Visibility |
| Eligible runs | Runs that completed and are relevant. Failed or unavailable runs are excluded, never counted as misses. | AI Visibility |
| Observation | One prompt run on one engine, with its raw and parsed forms. | AI Answers |
| Extraction confidence | How sure the parser is about each extracted field. | AI Answers |
| Coverage | The share of activated prompts with a successful observation per engine. | AI Answers |
| Citation rate | Runs with a brand domain citation divided by eligible runs. | Sources |
| Cited domains | Distinct domains cited across the selected runs. | Sources |
| Owned share | The share of citations that point to your own domains. | Sources |
| Share of voice | Brand mentions divided by total brand and competitor mentions in the same scope. | Competitors |
| Comparable slice | The same prompts, engines, market and window for every brand compared. | Competitors |
| Candidate | A discovered brand that is not yet confirmed as a competitor. | Competitors |
| Impressions and clicks | As reported by Google Search Console for the connected property. | Search |
| Position | Average position as reported by Search Console. | Search |
| Demand weight | Used to prioritise topics, not to change AI visibility scores. | Search |
| AI referral sessions | Sessions whose source matches an AI referral rule. | Business Impact |
| Conversions | GA4 conversion events for AI referred sessions. | Business Impact |
| Attribution window | The period used to connect visibility changes to outcomes. | Business Impact |
How Performance connects
When a number moves or looks wrong, Intelligence explains why.
Questions it answers
- How often do the AI engines name and recommend us?
- What changed since the last comparable run?
- Which competitors are named next to us?
- Which sources do the answers cite?
What it is not
- Explaining why
- Performance reports what is happening. Intelligence explains the cause.
- Guaranteeing a rank
- AI answers are not a ranked list, so results are rates over many runs.
- Reproducing the consumer app
- Engines are observed through provider APIs, which may differ from what a person sees in the app.
Performance questions
Which engines are measured at launch?
Gemini, Google AI Overview, Google AI Mode, ChatGPT and Perplexity. Claude, Copilot and Grok are coming in later versions.
Why show sample size on every score?
A rate from few runs is weak evidence. Showing the run count lets you judge how far to trust it.
Do I need Search Console or GA4?
No. They are optional and coming in V2. The first audit works without them.
See Performance on your own brand.
Start with a first audit. We show where you appear, who is named instead, and what to fix first.