AI Share of Voice (SoV) & Multi-Model Citation Audit

Evaluate brand presence, mention frequency, citation share of voice, and competitive positioning across generative answer engines (ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews) on core ICP queries.

Answer transcripts are supplied by the operator (pasted or exported). This skill does not query answer engines or third-party observation APIs on its own, and every figure it reports is a sample statistic over the corpus the operator supplied — not a platform-reported metric.

Transcript capture format, classification procedure, and the SoV computation live in references/checks.md.

Workflow

1. Define ICP Query Corpus & Competitor Cohort:
   • Assemble representative queries across 3 intent tiers:
     • Commercial / comparative: "best [category] software", "[brand] vs [competitor]".
     • Problem-solving / educational: "how to fix [problem]".
     • Transactional / feature-specific: "[feature] pricing and alternatives".
   • Identify 3–5 direct market competitors in the category. This cohort fixes the SoV denominator, so record it explicitly (see step 3).
   • Record corpus size, engine list, and capture date before computing any figure — see references/checks.md section 1 for the capture record format.