Included in Search and SEO skills
AI Visibility Review
v1.0.1A fixed prompt panel, per-provider observation ledger, inspected citation evidence and prioritised page actions with a repeatable retest plan.
Observe repeated ChatGPT answers and cited sources, then produce a defensible client report and specific page improvements to test.
What this Skill does—and does not do
ChatGPT only. No universal visibility score, causal attribution, guaranteed uplift, autonomous publication or customer-validation claim.
Inputs
- Brand, domain aliases, buyer task and market
- Competitors and a fixed branded/unbranded prompt panel
- Repeat count, source-research limits and total spend ceiling
Outputs
- Per-prompt and per-provider answer, mention, recommendation and citation ledger
- Coverage, failure and variability disclosure with sample denominators
- Inspected page/source recommendations and a repeatable retest plan
- Research receipt with exact run ids and costs
How this Skill works
- 1
Share the starting information
Give your agent the inputs listed above and explain the result you need.
- 2
Review the plan
Your agent reviews the proposed tools, inputs and maximum cost before any paid work begins.
- 3
Let your agent run the workflow
After you approve the plan, your agent uses the live Scrollport tools shown on this page and keeps the work inside the agreed limits.
- 4
Review the finished result
After approval, your agent returns the promised output and reports the tools used, final cost and any limitations.
Full instructions your agent receives
Give a business a reproducible record of what ChatGPT returned for selected buyer questions, the sources it cited and concrete improvements worth testing. The observations are a sampled baseline, not a measure of all customer exposure.
Use one authorised Scrollport connection with discover, inspect, run and wallet. Never call a supplier directly. This Skill measures ChatGPT only. Do not imply Google AI Overviews, Gemini, Claude or Perplexity were tested. Reuse sufficient supplied exports; when new calls are prohibited, work within that evidence and mark missing provenance rather than recollecting it.
Define the measurement
Read supplied product and customer context. Establish brand/domain aliases, market, buyer task, competitors, source pages, prompt count, repetitions and budget. Distinguish product lines, sibling brands and official domains before coding mentions. Ask only for missing choices that materially alter the report.
Default to three questions representing discovery, comparison and a concrete buyer problem, each repeated three times through one chosen collection route. The caller's panel and action limits override defaults; preserve supplied panels. Add a second route when the question includes collection consistency or the user needs a comparison; make that choice before collection, not after results. Record branded comparison prompts separately from unbranded discovery: a prompt that names the client is not an unbiased test of spontaneous brand visibility.
Freeze the exact prompt text, country, language and search settings before collection. Label prompts inferred from public information as analyst-selected; do not invent search volume or call them representative of real customer usage. Use actual customer questions or supplied search data when available.
Evidence routes and bounded cost
| Tool | Contribution | | --- | --- | | dataforseo.chatgpt-search | Observed answer and sources for an exact prompt | | brightdata.chatgpt-search
Cost and approvals
How pricing works
Fixed prompt repetitions on one collection route, optional route comparison and bounded source research.
Your approval points
- Use existing authorisation for the business, market, bounded selection rules and total maximum; obtain missing approval before spending
- Pause for a server confirmation request or a material scope or budget expansion
- No publishing or website writes without explicit scope