AI SEARCH AUDIT CHECKLIST
OppAlerts | September 10, 2026

PROJECT
[ ] Define the customer question and the decision the test supports.
[ ] Record exact prompts, versions, products, models, and search conditions.
[ ] Separate branded verification from unprompted discovery.
[ ] Set scoring rules before reviewing answers.

COLLECTION
[ ] Record every attempt, including failures and retries.
[ ] Save complete answers and available source/tool records.
[ ] Record account context, location, date, and settings where available.
[ ] Mark unavailable settings or tool records as unknown.

SCORING
[ ] Count each brand at most once per completed answer for each metric.
[ ] Separate mentions, recommendations, and links to the business domain.
[ ] Check factual claims against approved business information.
[ ] Resolve ambiguous names and scoring disagreements.
[ ] Report counts alongside rates and failed-attempt totals.

ACCESS AND CONTENT
[ ] Check important URLs, status codes, redirects, and returned content.
[ ] Review robots.txt and relevant verified crawler requests.
[ ] Review Google indexing, preview controls, and Search generative AI settings.
[ ] Compare visible facts, supported markup, feeds, and business profiles.
[ ] Keep prices, conditions, service areas, and product identifiers together.

CORRECTIONS AND FOLLOW-UP
[ ] Assign each correction an owner and acceptance check.
[ ] Save old/new content and publication dates.
[ ] Repeat comparable prompts after the correction.
[ ] Record model, source, account, and scoring changes.
[ ] Separate verified corrections from possible effects on recommendations.

Full method:
https://oppalerts.com/AI-Search-Optimization-Guide/AI-Visibility-Audit