Running your first AI visibility audit
An audit is a dated diagnostic snapshot. It helps you inspect your brand's visibility and choose work to investigate. Its score is only meaningful alongside scope, coverage, and evidence.
What the audit covers
The current full audit covers five eligible areas:
- AI presence: Checks whether your brand appears in the sampled AI answers.
- Entities: Checks brand identity and consistency across available sources.
- Reviews: Checks applicable review platforms and available feedback.
- On-page: Reviews website content and structure.
- Citations: Looks for relevant sources and brand mentions.
Press is coming soon and excluded from the score. Completion time depends on the selected scope and source availability; do not promise a fixed turnaround.
Step-by-step walkthrough
Step 1: check the brand setup
In an authorized workspace, select the brand you intend to audit. Review its website, brand brief, active keywords, and competitor names. Research suggestions need human review before they become a reliable brief.
Step 2: choose the scope
- Full: All five currently eligible areas. Press is excluded.
- Quick: Citations, AI presence, and on-page. Entities and reviews are outside this scope.
- AI presence only: Sampled AI answers.
- Citations and AI presence: Those two areas together.
Choose the scope for the question you need to answer. A public free scan is narrower than the workspace full audit; do not present it as the same report.
Step 3: check completion and coverage
Review the audit status and any incomplete, failed, or excluded areas. A missing measurement is not a measured zero. Wait for sufficient evidence before drawing conclusions or sharing a report.
Step 4: read the findings
Composite score: A weighted score over the declared eligible scope. Scope exclusions can change applied weights, but failed checks do not redistribute their weight. Incomplete evidence may produce a partial lower-bound score or an unavailable score. Compare snapshots with the same scope, coverage, and score basis.
Area breakdown: Inspect each finding and its underlying sources. A low score is a reason to investigate, not proof that one tactic will deliver the greatest return.
Summary and suggested actions: Review the generated explanation for factual accuracy, relevance, and feasibility. Treat impact estimates as suggestions, not measured outcomes.
Interpreting common patterns
Sparse findings: Check keyword relevance, source coverage, and missing data before assuming the brand has no public presence.
Strong reviews, weak sampled AI presence: Read the AI answers and cited sources. Review quality alone does not establish whether an engine will recommend the brand.
Public mentions, weak sampled AI presence: A mention may not be indexed, retrieved, or relevant to the prompts being tested. Investigate rather than assuming a delay will resolve itself.
Sharing an audit with a prospect
Explain the date, scope, tested prompts, and coverage. Use actual findings rather than invented competitor scores. Separate a proposed work plan from an outcome promise, and be clear about the difference between the public free scan and a workspace audit.
Re-running audits
Save a baseline, then choose regular review checkpoints that suit the work. Keep scope and coverage comparable, and record changes to the prompt set or business setup. A higher visibility score alone does not establish revenue return or show which action caused a change.
In Outrigger
Use the Audits module in your authorized workspace to review status, findings, and available reports. Check the evidence and exclusions before sharing an export.