AI Visibility Audit

This skill answers two questions: where does the brand stand in AI answers today, and what is holding it back? It baselines real engine responses, scores readiness across five dimensions, and returns a prioritized roadmap. It is the diagnostic that points every other AEO skill at the right work. It applies to any business, B2B or B2C.

When to use this skill

• Running an initial AEO/GEO audit for a brand or client.
• Diagnosing why a brand is absent or misrepresented in AI answers.
• Producing a prioritized AEO roadmap.
• Re-auditing after major content, technical, or authority changes.

Before you start

Read aeo-foundation.md for the brand, the prompt set, and competitors. If the prompt set is thin, run prompt-research first — you cannot baseline without prompts. If the foundation is missing, point to the aeo-foundation skill.

Step 1 — Baseline the current state

Run the prompt set across the major engines and record, per prompt and per engine:

• Is the brand cited (linked), mentioned (named, no link), or absent?
• Who is cited instead, and in what order?
• What is the sentiment and is anything factually wrong?

This snapshot is both the diagnosis input and the measurement starting line.

Step 2 — Score the five dimensions

Apply the shared rubric in ../../references/aeo-audit-rubric.md, scoring each 0–3:

1. Discoverability — can engines reach the content? (crawler access, rendering, speed, indexable pages)
2. Interpretability — can they understand it? (structure, answer-first, schema, entity clarity)
3. Attribution — will they cite it? (named authors, sourcing, original data)
4. Authority & trust — will they trust it? (third-party corroboration, reviews, freshness)
5. Governance & risk — is it safe to cite? (accuracy, compliance, consistent facts)

Step 3 — Diagnose

For each prompt the brand loses, trace the cause to a dimension and a lever:

• Absent everywhere → usually discoverability or entity problems.
• Ranks/known but not quoted → interpretability (structure/extractability).
• Quoted less than competitors → authority (corroboration, freshness, reviews).
• Misrepresented → governance/entity consistency.

Step 4 — Prioritize and deliver

Sort fixes by impact vs. effort into critical (blockers), high-impact, and long-term, and map each to the skill that executes it. Use templates/audit-worksheet.md to capture scores and the fix list, and map the result to a maturity stage (Invisible → Emerging → Established → Authority).

Output format

• Baseline table — presence, competitors, sentiment per top prompt and engine.
• Dimension scores — 0–3 each, /15 total, with evidence.
• Prioritized fixes — issue, why it matters, the fix, the owning skill, rough effort.
• Maturity stage and the few moves that advance it.
• (assumption) tags where access or data was limited.

Common pitfalls

• Scoring structure without first baselining real answers.
• A 200-item checklist no one will action — prioritize.
• Auditing one engine and generalizing; they differ (see the platform skills).
• Treating the score as the deliverable instead of the roadmap.

Tool

Score the five-dimension rubric and get a maturity stage:

python scripts/score_audit.py --template          # blank input to fill in
python scripts/score_audit.py --input scores.json # total /15 + stage + weakest dimensions

For a fast manual pass, use the shared AEO readiness checklist.

Optional. The script only automates this step; it needs python3 (standard library only, no install). No Python? Just follow the steps above manually.

Related Skills

• aeo-foundation / prompt-research — supply the prompts to baseline against.
• competitor-ai-analysis — quantifies share of voice vs. rivals.
• ai-visibility-tracking — turns this baseline into ongoing measurement.
• Content, authority, technical, and platform skills — execute the prioritized fixes.