SpyderBot Documentation

AI Visibility Benchmark

AI Visibility Benchmark is a structured, longitudinal measurement workflow built from Prompt Universe.

What is AI Visibility Benchmark?

It helps you repeatedly measure strategic AI Visibility using a stable, reviewable set of questions and controlled contexts.

The workflow is:

Prompt Universe
→ Build AI Visibility Benchmark
→ Review
→ Start Observing
→ Prompt Observatory

You do not need to run Prompt Explorer before starting a Benchmark.


Benchmark measurement classes

The Benchmark uses three classes.

Discovery Visibility

Measures whether your Brand and competitors appear in Brand-neutral discovery questions.

Brand Understanding

Measures how AI systems describe and evaluate the Brand using controlled, comparable Brand contexts.

Competitive Capture

Measures competitive behavior using comparator-specific controlled contexts.


Core vs Supplemental

Benchmark candidates may be classified as:

Core
Supplemental

Only eligible Core membership contributes to headline Benchmark metrics.

Supplemental items provide diagnostic context without silently changing headline methodology.


Headline metrics

The Benchmark uses exactly four headline metrics:

Discovery Visibility
Controlled AI Sentiment
Competitive Capture
Visibility Gap

There is no required overall Benchmark score.


Controlled AI Sentiment

Controlled AI Sentiment is a Benchmark metric derived from eligible controlled Brand Understanding contexts.

It is not the same as general AI Brand Portrayal sentiment.

The two may reuse underlying response classification, but they represent different measurement scopes.


Benchmark history

Benchmark history begins from the active Benchmark measurement version.

Changing Benchmark membership or a material measurement scope should not silently rewrite historical results.

When a newer configuration is needed, SpyderBot may create or propose a new Benchmark version for review.


Reading Benchmark results

Use:

  • Discovery Visibility to understand discovery presence;
  • Controlled AI Sentiment to understand controlled Brand evaluation;
  • Competitive Capture to understand comparator-specific competitive performance;
  • Visibility Gap to understand the difference between your Brand and the selected comparator.

Longitudinal charts use actual eligible observation dates.

Missing observations are not treated as zero and are not silently carried forward.


Next step

When you choose Start Observing, continue with Benchmark Observatories.