SpyderBot Documentation

Visibility and competitive presence

AI Visibility describes a broad area of measurement: how brands appear and are represented in AI answers. The percentage named Organic Visibility measures one part of that area. It is not a combined score of sentiment, recommendations, citations and traffic.

The definitions below apply to the current measurement method. A historical result keeps its original method and scope. All examples on this page are fictional illustrations, not customer observations or evidence-quality guarantees.

Response rates are balanced within their declared measurement scope. Read How aggregation works ↗ before comparing the displayed percentage with raw counts.

On this page

Organic Visibility · Aided Representation · Mention Count · AI Share of Voice · Observed Prompt Reach · Observed AI System Presence · Visibility by AI System · Visibility and Recommendation Gaps · Competitive Capture

Organic Visibility

Unit: Percent

How often your brand is substantively mentioned in answers to eligible open discovery questions.

How it is calculated

For each eligible prompt and AI-system configuration, calculate the share of complete, usable answers that substantively mention the target. Combine these results using the declared balanced aggregation policy. Core Prompts have equal strategic weight under the default profile; wording variants and additional repetitions do not create extra strategic votes.

Open discovery questions do not preselect brands in a way that turns the task into aided evaluation, named comparison or a competitor-led alternative search. A complete eligible answer without a substantive target mention contributes no mention. Missing collection or unresolved extraction does not become a negative answer.

How to interpret it

A mention can be positive, neutral or negative. This percentage does not measure recommendation, truth, reach among all AI users or real-world market share. A product or retailer mention does not automatically award its parent brand a mention.

What to do next

Inspect the prompts and systems with low visibility, then check recognition and relevance before planning content changes.

Illustrative example

In one complete, equally weighted two-system example, system A mentions a brand in 8 of 10 answers and system B in 2 of 2. Their balanced visibility is (80% + 100%) / 2 = 90%. The pooled 10/12 = 83.3% answers a different question. A full report also balances its declared prompt scope.

Related concepts: Aided Representation ↗ · AI Share of Voice ↗ · Global Organic Visibility Exposure ↗

Measurement scope and aggregation ↗ · Evidence and availability ↗

Aided Representation

Unit: Percent

How often an answer substantively discusses the target in eligible questions that already introduce it.

How it is calculated

Substantive target discussion is counted within eligible aided-evaluation answers and combined using balanced aggregation. Merely repeating the supplied name is not substantive discussion. Aided results have their own population and do not enter organic discovery.

How to interpret it

Aided Representation answers what happens when the brand is introduced. It is not evidence that AI would discover the brand without that introduction.

What to do next

Review portrayal, claims and endorsement within the same aided context.

Related concepts: Organic Visibility ↗ · Branded Endorsement Rate ↗

Measurement scope and aggregation ↗ · Evidence and availability ↗

Mention Count

Unit: Observed response–entity count

The number of distinct collected answer artifacts that substantively mention the selected entity in the displayed scope.

How it is calculated

Count each substantive entity mention once per distinct accepted physical answer artifact. Repeated spellings or repeated mentions within that answer do not add more mention units. A reused artifact is not another independent answer.

How to interpret it

This is an evidence-volume count. It is not the numerator of a balanced headline percentage. Counts depend on collection depth and scope.

What to do next

Use visibility rates for comparable performance and counts to understand available support.

Illustrative example

An answer containing the brand name three times contributes one mention unit for that brand.

Measurement scope and aggregation ↗ · Evidence and availability ↗

AI Share of Voice

Unit: Percent of tracked exposure

Your share of balanced brand exposure within the declared tracked brand set.

How it is calculated

Calculate each tracked brand’s balanced visibility on the same eligible observations. Divide the target’s exposure by the sum of those brand exposures and multiply by 100. All brands use the same scope and weights. If total exposure is zero, share is unavailable.

How to interpret it

The shares sum to 100% across the declared set. Individual visibility rates need not sum to 100% because one answer can mention several brands. Adding or removing a competitor changes the comparison basis.

What to do next

Review both your own visibility and competitor exposure before interpreting a share change.

Illustrative example

Balanced exposure of 60%, 30% and 10% gives shares of 60%, 30% and 10%. If the exposures are 60%, 60% and 30%, the shares are 40%, 40% and 20%.

Related concepts: Visibility and Recommendation Gaps ↗ · Organic Visibility ↗

Measurement scope and aggregation ↗ · Evidence and availability ↗

Observed Prompt Reach

Unit: Percent of eligible prompt units

The share of eligible Core Prompt units with at least one observed brand mention at the displayed sampling depth.

How it is calculated

Count fully observed eligible prompt units with at least one substantive target mention, divided by the fully observed eligible prompt units in the declared panel. A comparable full-panel headline requires the planned depth and system/surface coverage for the entire panel. With incomplete coverage, any descriptive subset is explicitly labelled and the original planned total remains visible.

How to interpret it

More observation opportunities can increase at-least-once reach without improving the underlying mention rate. A missing prompt is not a negative observation.

What to do next

Compare reach only at compatible depth and scope; use the per-prompt visibility rates to investigate concentration.

Measurement scope and aggregation ↗ · Evidence and availability ↗

Observed AI System Presence

Unit: Percent of measured systems

The share of measured independent AI systems with at least one brand mention in the complete displayed panel.

How it is calculated

The number of fully observed systems containing at least one substantive mention is divided by the number of systems in the complete declared panel. The system count and sampling depth explain the opportunity behind this percentage.

How to interpret it

This describes at-least-once presence, not average visibility and not worldwide platform usage. Several surfaces of the same underlying system do not automatically represent independent systems.

What to do next

Open the system breakdown to distinguish a single mention from consistent visibility.

Measurement scope and aggregation ↗ · Evidence and availability ↗

Visibility by AI System

Unit: Percent

The selected visibility metric calculated separately for each actual AI system.

How it is calculated

Project the same eligible metric and balanced prompt scope onto each actual system. Do not pool more heavily sampled prompts into a larger strategic weight. Compare systems on compatible prompt support.

How to interpret it

A missing or unknown system result is not zero. A product name and an exposed model identity are different; an undisclosed model version stays unknown.

What to do next

Investigate material differences on the same questions and time window.

Measurement scope and aggregation ↗ · Evidence and availability ↗

Visibility and Recommendation Gaps

Unit: Percentage points

The difference between your result and the stated competitor’s result on the explicitly named visibility or recommendation metric.

How it is calculated

Gap = your percentage minus the stated comparator’s percentage, on the same metric and common eligible observations. A Visibility Gap uses visibility; a Recommendation Gap uses active recommendation. A benchmark time series keeps its selected comparator fixed. A card explicitly labelled “Gap vs best tracked competitor” uses the best competitor in the declared current roster and preserves ties. These comparator policies are different.

How to interpret it

Positive means ahead and negative means behind on the named metric. The gap is in percentage points, not relative percent growth or statistical significance. A changing comparator can change the meaning of a series.

What to do next

Check the named metric, common question scope and comparator before investigating the supporting answers.

Illustrative example

Your visibility is 45%, the stated competitor is 60%: the gap is −15 percentage points, not −15% relative change.

Measurement scope and aggregation ↗ · Evidence and availability ↗

Competitive Capture

Unit: Percent of competitor-led alternative answers

How often the target is substantively represented in questions asking for alternatives to a declared competitor.

How it is calculated

Use eligible alternative-seeking questions with the stated competitor anchor. Count substantive target representation across all usable eligible answers, then apply balanced aggregation within that anchor-specific cohort. The target need not be actively recommended to qualify as represented.

How to interpret it

This is distinct from organic discovery and from Alternative Recommendation Rate. A named competitor already shapes the opportunity.

What to do next

Review where the target appears as an alternative and whether it is actually recommended.

Related concepts: Alternative Recommendation Rate ↗ · Organic Visibility ↗

Measurement scope and aggregation ↗ · Evidence and availability ↗