AI Visibility Report: What Search Console Can Tell You

Google Search Console includes AI Overviews and AI Mode in its Web search totals. Those totals alone do not isolate AI traffic. An AI Overview’s links share a search position, not a citation rank. Use Search Console alongside on-site outcomes and separately documented observations.
The short version
Google Search Console can supply Google Search performance data for an AI visibility report, but its Web totals alone cannot tell you how much came from AI answers. Google’s guide to AI features says traffic from AI Overviews and AI Mode is included in the Performance report’s Web search type.
The practical distinction is between a reported measurement and an explanation for that measurement. A change in Web clicks is a measurement. Saying an AI answer caused it requires evidence beyond that total.
Web totals are not AI-only attribution
Read the Web report as a combined view of Google Search performance, not as a count of AI recommendations. The inclusion of AI features does not make the entire total attributable to those features.
For example, if clicks to your mortgage page rise, the Web total establishes that the page received more reported search clicks. It does not, by itself, establish that AI Overviews drove the increase or that Google recommended your institution more often.
Recommendation: Label this section of your report “Google Search performance, including AI features.” Keep any explanation about AI contribution separate from the measured totals.
Also keep the reporting boundary explicit. Search Console Help says Search Console reports visibility and interactions in Google Search, News, and Discover. These measurements are not evidence of your institution’s visibility across other assistants.
AI Overview position is not citation rank
An AI Overview occupies one position in Google Search results. Every link within that overview receives the same position, according to Google’s position definitions.
That means Search Console position does not tell you which institution was cited first inside the answer. It does not establish that your link appeared before a competitor’s link within the overview.
There is another distinction to preserve. Search Console’s average position uses the topmost position of a site’s or page’s link, averaged across queries where it appeared. A link needs an impression for its position to be recorded.
Recommendation: Retain the label “average position.” Do not rename it “AI citation rank” or use it as a measure of prominence within an AI answer.
For AI Overviews, an external link click counts as a click, and standard impression rules apply. Those definitions describe recorded exposure and interaction, not whether the answer endorsed your institution.
On-site outcomes answer a different question
Search performance tells you about visibility and interaction in search. It does not complete the account of what visitors did after arriving.
Google recommends tracking conversions and time spent on site in tools such as Google Analytics alongside Search Console.
Recommendation: Place relevant on-site outcomes beside search performance, with separate labels. For a bank marketer, that might mean reviewing an application-start event or appointment request if your institution already measures it.
Do not label those outcomes “AI conversions” merely because the landing page also appeared in an observed AI answer. Parallel movement is a reason to investigate, not proof of attribution.
A reporting note bank marketers can reuse
Recommendation: Add this interpretation note beneath your Google Search figures:
Google Search Console Web totals include traffic from AI Overviews and AI Mode. These totals do not independently establish AI-only traffic or recommendation frequency. AI Overview links share a search-result position, so average position is not citation order. On-site outcomes are reported separately and are not attributed to AI without supporting evidence.
This note works for a product-page report or an institution-wide review. For audience-specific context, continue to the pages for community bank marketers or credit union marketers.
Keep manual observations separate
A screenshot can document what you saw. It should not be treated as an explanation for an aggregate Search Console result.
Google explains that AI Mode and AI Overviews may use different models and techniques, so their responses and links vary. AI Overviews also often do not trigger. Search Console Help notes that location and search history can affect results, and that Search Labs experiments are excluded from its data.
Recommendation: Record the query, observation date, location context, AI feature, and cited URL when checking manually. Keep those observations in a separate section from Search Console measurements.
For your next report, separate three kinds of evidence: Google Search totals, observed AI appearances, and on-site outcomes. For each claimed improvement, identify which evidence supports it and which attribution remains unknown.
Questions people ask
Can I report all Search Console Web clicks as AI traffic?
No. Google’s guide to AI features says AI Overviews and AI Mode are included in Web totals, but those totals alone do not isolate their contribution. Label the figures as Google Search performance, including AI features.
Does average position show where my bank was cited inside an AI Overview?
No. Google’s position definitions assign the same search-result position to every link within an AI Overview. That position does not measure citation order inside the answer.
Why does my manual search look different from the report?
Google’s Search Console documentation says individual results can differ because of factors such as location and search history. Google also says AI Overviews and AI Mode can produce different responses and links, so treat a manual check as an observation rather than a reproduction of aggregate data.
What should I put beside Search Console figures?
Google recommends reviewing conversions and time spent on site in tools such as Google Analytics. Keep those outcomes separately labeled, and do not attribute them to AI solely because they changed alongside search performance.