Measurement / framework / Free to use
AI Search Impression Review: A Scope and Evidence Checklist
A practical worksheet for comparing Google AI-feature and overall Web impressions without mixing scopes, sources or business outcomes.
Growthcraft Editorial · 2026-10-29. AI-assisted research and implementation. Examples are synthetic; Akshay's personal review is not claimed.
Copyable template
Select and copy the complete template below. With JavaScript enabled, you can edit, copy and download it in the interactive workspace.
AI SEARCH REPORT CONTRACT — original Growthcraft worksheet Decision and owner: Google Search property or exact canonical page: Finalized start and end dates (Pacific Time): Search type (Web text or Web multimodal): Country and device filters: Aggregation: property chart OR same single page; do not mix them. AI-feature impressions: exact value / not available / preliminary. Overall Web impressions: exact value / not available / preliminary. Extraction time and report source for both: Are all scope fields identical? yes / no / unknown. Are both counts exact and finalized? yes / no / unknown. If either answer is no or unknown, do not calculate a share. If yes and overall Web impressions are positive: 100 × AI impressions / Web impressions. If both are zero: share undefined. If AI exceeds Web: investigate mismatched data. Page table rows can overlap at property level: do not add them to recreate a site total. Google's AI figure is included within the Web figure: never add them together. Separately record Bing citation counts, referral sessions, actual enquiries and their scopes. Next action, evidence owner and a comparable review window: Limitations: this does not identify a person, search prompt, click, recommendation or sale.
What decision does this support?
Use this original worksheet when a team wants to understand how its site appears in Google's generative Search features while retaining a separate view of traffic and enquiries. It is a reporting contract, not a method for finding a person's prompt or predicting which pages an AI will recommend.
Four checks before a ratio
- Scope: select the same property or exact page, time window, Web search type and country/device filters.
- Aggregation: use matching property-level chart totals or the same exact page. Do not sum page rows and compare that sum with a deduplicated property chart.
- Completeness: record the extraction time and whether both reports are final; unavailable, suppressed and preliminary are not measured zero.
- Decision: if the evidence is compatible, compute a share; keep citations from other systems, referral visits and enquiries separate.
Worked synthetic example
In one finalized 28-day window, a property has 250 AI-feature impressions and 2,000 total Web impressions. With all scope conditions aligned, 250 ÷ 2,000 × 100 = 12.5%. The 250 is already included in the 2,000; adding them would double-count. This says nothing about clicks or unique people. Try the calculator.
Assign the follow-up
An analyst records sources, filters and counts; an editor reviews the cited or appearing pages for usefulness; a sales owner examines genuine enquiries under a separate definition. Recheck equivalent complete windows without attributing changes to an edit merely because it happened first.
When not to use the ratio
Stop if scope differs, data is incomplete, the denominator is zero, values are rounded, or AI exceeds the supposed total. Bing's citations, ChatGPT mentions and Google AI impressions are not interchangeable. Do not change crawler access solely to make this metric move.
Sources and boundaries
Checked 29 October 2026. Google's generative AI Search report documentation describes the report, supported features, aggregation and export caveats. Microsoft's 13 October 2026 Topics Portfolio announcement explains a different, approximate beta report based on grounding queries. These are different providers and measurements; neither is a count of leads or a guarantee of future visibility. The worksheet and numerical examples are original, synthetic editorial material.