SpyderBot Documentation

GEO Audit and outcome verification

GEO Audit diagnoses a site’s observed readiness and representation. Diagnostic findings and remediation progress are separate from measured AI visibility, citations or traffic. Fixing a technical issue does not guarantee that an AI system will mention or recommend a brand.

On this page

LLM Readability · URL Optimization · AI Entity Representation · Remediation Verification Status

LLM Readability

Unit: Findings and versioned diagnostic result

Evidence about access, discovery, renderability, extraction, readability and content structure under the tested conditions.

How it is assessed

LLM Readability evaluates the content available under the audit’s tested conditions. Its findings cover five related questions:

Assessment groupWhat it tells you
AccessWhether the tested resource can be reached and its response delivered.
DiscoveryWhether relevant resources can be found through the discovery paths examined by the audit.
RenderabilityWhether the relevant content is present in the tested delivery or rendering state.
Extraction and readabilityWhether the main information can be extracted and read from that tested content.
Content structureWhether the organisation of the content makes its information understandable.

A finding is based on the checks actually performed on the tested pages and time. Unavailable or excluded checks are evidence limitations, not successful checks. A displayed diagnostic score summarises findings under the audit method version; it is not a simple percentage of all possible website checks. The scoring implementation remains private. The finding-level evidence is the useful basis for deciding what to fix.

How to interpret it

A probe is not a real bot revisit, indexing confirmation or proof of training access.

What to do next

Verify configuration and observed delivery, then monitor real activity.

Related concepts: Remediation Verification Status ↗ · Successful Access Rate ↗

Measurement scope and aggregation ↗ · Evidence and availability ↗

URL Optimization

Unit: Findings and versioned diagnostic result

Evidence about tested URL delivery and identity, including status, redirects, canonicalisation, parameters, duplicates, locale routing and access/indexability-related configuration.

How it is assessed

URL Optimization assesses the identity and delivery of the tested URLs:

Assessment groupWhat it tells you
HTTP statusThe response outcome observed for the tested URL.
Redirect behaviourHow the tested request moves to its destination.
CanonicalisationWhich URL is presented as the preferred identity and whether the observed signals agree.
Parameters and duplicatesWhether alternate URL forms create ambiguity or unnecessary duplication in the tested scope.
Locale routingWhether the tested language or regional route reaches the intended content.
Access and indexability-related configurationWhether the examined configuration presents an obstacle under the tested conditions.

Findings retain the tested URL, evidence and audit version. Unobserved conditions remain unknown. AI Traffic can reveal a reason to investigate, but traffic activity does not automatically change an audit score. A rerun creates a new audit result, so you can compare the actual observations before and after a change.

How to interpret it

A URL diagnostic is not a predicted citation probability or guaranteed ranking.

What to do next

Address the supported URL issue, verify delivery and evaluate subsequent observed outcomes.

Related concepts: Remediation Verification Status ↗

Measurement scope and aggregation ↗ · Evidence and availability ↗

AI Entity Representation

Unit: Findings and versioned diagnostic result

Evidence about how clearly the tested content represents the intended organisation, product or person.

How it is assessed

AI Entity Representation examines how the tested content describes the intended entity. Earlier product references may call this Entity Understanding.

Assessment groupWhat it tells you
Entity identityWhether names and identifying descriptions distinguish the intended organisation, product or person.
Structured factsWhich explicit facts about the entity appear in the tested content.
Schema consistencyWhether structured information agrees with the other examined representations.
Entity relationshipsHow the tested content connects the entity with relevant organisations, people, products or services.
Conflicting claimsWhere the examined evidence gives inconsistent or ambiguous accounts.

The assessment uses the available site facts, schema and linked representation evidence in its declared scope. Bot requests alone cannot show that an AI system misunderstands an entity. Findings and diagnostic summaries belong to the identified audit method; they are not a view into a model’s hidden knowledge graph or a universal entity-accuracy percentage.

How to interpret it

The diagnostic does not reveal a model’s hidden knowledge graph or guarantee correct future identification.

What to do next

Review naming, relationships and supporting factual content.

Related concepts: Remediation Verification Status ↗ · Founder Mention Coverage ↗

Measurement scope and aggregation ↗ · Evidence and availability ↗

Remediation Verification Status

Unit: Evidence milestones

The stage of evidence supporting a claimed fix and its subsequent observed outcome.

How it is assessed

Each status corresponds to a different evidence milestone: a fix was reported; configuration was checked; a probe succeeded; real bot activity revisited the page; a comparable outcome was evaluated; or a regression was observed. A milestone applies only when that type of evidence exists. A successful probe alone does not establish a real bot revisit or outcome improvement.

How to interpret it

Verification is not a causal claim that the fix produced an AI answer change.

What to do next

Open the evidence for the current milestone and continue monitoring.

Measurement scope and aggregation ↗ · Evidence and availability ↗