Dashboard Metrics
These metrics summarize the scope of your monitoring.
Prompts
What it measures
The total number of active prompts being monitored.
How it’s calculated
Counts all active custom prompt variants (not Seed Prompts) that match your selected filters (persona, tags, topics, date range, etc.).
Responses
What it measures
The total number of AI responses collected during the selected time period.
How it’s calculated
Counts distinct observations that:
Are linked to an active custom prompt variant
Fall within the selected date range
Come from a supported AI platform
Match applied filters (persona, tags, platform, etc.)
Platforms
What it shows
Which AI platforms produced responses during the selected period.
How it’s determined
Identifies all unique AI assistants present in observations matching your filters.
Time Series Metrics
These metrics are shown over time and automatically aggregated based on the selected date range.
Brand Presence
What it measures
The percentage of AI responses where your brand is mentioned.
How it’s calculated
For each time period:
Total observations = all responses
Present observations = responses where your brand is detected
Formula:
(Present / Total) × 100
Time aggregation
0 to 14 days: Daily
15 to 90 days: Weekly
90+ days: Monthly
Competitive Presence
What it measures
How often your brand and competitors appear in AI responses over time.
How it’s calculated
Brand presence: observations where is_brand_present = TRUE
Competitor presence: observations where each competitor is detected
Denominator for both is total observations for your prompts
Each brand or competitor is shown as its own time series of raw counts.
Time aggregation
Same as Brand Presence.
Position
What it measures
Where your brand appears in AI responses.
Positions
Top: mentioned in the top 25% of the response
Middle: mentioned in the middle 50% of the response
Bottom: mentioned in the bottom 25% of the response
Missing: position not detected
How it’s calculated
Percentages are calculated using only observations where your brand is present:
Top % = Top / Total excluding missing
Middle % = Middle / Total excluding missing
Bottom % = Bottom / Total excluding missing
Position metrics only include responses where your brand is present.
Sentiment
What it measures
How AI describes your brand.
Sentiment types
Positive
Mixed
Negative
Missing
How it’s calculated
Percentages are calculated using only observations where sentiment is available:
Positive % = Positive / Total excluding missing
Mixed % = Mixed / Total excluding missing
Negative % = Negative / Total excluding missing
Sentiment metrics only include responses where your brand is present.
Citations
What it measures
How often your brand is cited as a source, compared to competitors and third parties.
Citation types
Brand: your registered domains from the Context tab
Competitor: competitor domains
Other: third-party sources
How it’s calculated
Each citation URL is classified by owner
Counts are aggregated by time period
Brand citation % =
(Brand citation observations / Total observations) × 100
Time aggregation
Daily, weekly, or monthly depending on date range.
Time Ranges and Aggregation
Default range
Last 90 days.
Automatic aggregation
0 to 14 days: Daily
15 to 90 days: Weekly
91+ days: Monthly
This keeps charts readable while preserving trend accuracy.
Filter Behavior
All metrics respect applied filters, including:
Persona
Tags
Topics
Platform
Category
Branded vs non-branded
Favorites
Date range
Brand presence
Competitor presence
Notes and Caveats
Metrics are calculated in real time when the dashboard loads or filters change
“Missing” values appear when position or sentiment cannot be determined
Only AI platforms marked as visible for your brand are included
If something looks unexpected, contact support with the metric name and date range you’re reviewing.
Citations Tab Metrics
These metrics explain how sources are cited by AI assistants and how citation influence is calculated across URLs and domains.
Main Table Metrics
The citations table can be grouped by Domain or URL. Each row represents a single source and includes the following metrics.
Prompts
What it measures
The number of unique prompts that have cited this source at least once during the selected time period.
How it’s calculated
Count of distinct prompt IDs where at least one response cited the source.
Formula:
COUNT(DISTINCT prompt_id)
Responses
What it measures
The total number of AI responses that cited this source during the selected time period.
How it’s calculated
Count of distinct observations that include a citation to the source.
Formula:
COUNT(DISTINCT observation_id)
Citation Consistency
What it measures
How consistently a source is cited across AI responses.
How it’s calculated
(Number of responses that cited this source ÷ Total number of responses that cited any source) × 100
Influence Score
What it measures
A composite score that reflects both how often a source is cited and how broadly it appears across prompts.
How it’s calculated
Citation Consistency (%) × Number of Unique Prompts
This favors sources that are cited frequently and across many different prompts.
Owner Type Classification
Each source is classified into one of three ownership categories.
Owner Types
Brand
Sources owned by your brand, matched against configured brand websites.
Competitor
Sources owned by configured competitors.
Other (Third party)
Sources not owned by your brand or competitors.
Domain Ownership Logic
Ownership is determined using a hierarchical matching process:
URL-level match for brand websites
URL-level match for competitor websites
Domain match for primary brand website
Domain match for alternative brand websites
Domain match for competitor websites
Default to third party if no match is found
Conservative Domain Classification
When grouping by domain, all URLs within a domain must match the same owner type.
Logic:
If all URLs in a domain have the same owner type, the domain is assigned that owner.
If URLs map to multiple owners, the domain is classified as Third party.
This prevents over-attribution on shared platforms like social networks or publishing sites.
Domain-Level Visualizations
Top Domains Cited
What it shows
The top five cited domains by total response count.
How it’s calculated
Domains are ranked by total citation volume. All remaining domains are grouped into an Other category.
Formula:
SUM(response_count) grouped by domain
Citations by Owner
What it shows
The percentage breakdown of citations by owner type (brand, competitor, third party).
How it’s calculated
For each owner type:
(Owner citation count ÷ Total citation count) × 100
Percentages are rounded so the final values sum to 100%.
Owner Time Series
What it shows
How brand, competitor, and third-party citation percentages change over time.
How it’s calculated
For each time period:
(Owner citation count for the period ÷ Total citation count for the period) × 100
Additional Metadata
Brand and Competitor Mentions
Indicates whether the source content explicitly mentions your brand or competitors.
Topics
Displays topics associated with the cited source or domain.
Mixed Ownership Indicators
For domains containing both brand-owned and third-party URLs, multiple owner labels are shown with explanatory tooltips.
Notes
All citation metrics respect applied filters (date range, persona, topic, tags, platform, etc.)
Citations support both URL-level and domain-level grouping with drill-down
Time aggregation (daily, weekly, monthly) adjusts automatically based on the selected date range
Summary metadata (top domains and owner breakdowns) is calculated independently to remain consistent across pagination