Google Search Console Generative AI Report

What Google Search Console’s Generative AI Report Shows Your Business

Google Search Console now gives website owners a dedicated view of their visibility in Google’s generative AI search features. Google announced the reports on June 3, 2026, and says the worldwide rollout was completed on August 31. Separate reports cover Search and Discover. (Google)

If AI impressions rise while inquiries stay flat, what should your team investigate? For a business in Houston or The Woodlands, start by identifying which pages received recorded exposure in Google’s AI search experiences and which customer questions those pages answer.

Exposure does not establish customer acquisition. Use the report to identify pages worth investigating, then compare that visibility with separate traffic and inquiry evidence.

This guide explains the current report, provides a worked example, and shows how to turn it into a monthly review. It is designed to fit within a broader SEO reporting process, with clear definitions and practical next actions.

What the Report Shows in October 2026

The dedicated Search report documents impressions for AI Overviews and AI Mode. Its controls include pages, countries, dates, devices, and text-based versus multimodal search. It does not document a query dimension or dedicated clicks, CTR, average position, city, or ZIP-code fields. It does not provide a documented AI Overviews-versus-AI Mode breakdown. (Google)

A frequently shown page may answer a buying question, a broad research question, or a topic only loosely related to your services. The report identifies exposure; your team must assess its business relevance.

Google also says this AI data remains within the overall performance reporting. The dedicated report is a separate view of included activity. Do not add its impressions to your total Search impressions and present the sum as additional reach. (Google)

An Impression Is an Exposure Measure

Google’s measurement guidance describes impressions in terms of links being seen or potentially seen, with the rules depending on the result format. External links clicked in AI Overviews or AI Mode count as clicks in Search Console’s wider measurement system. That does not create a separate click field in this dedicated report. (Google)

Consider a fictional service page with 800 AI impressions during a review period. The defensible statement is that the page received 800 recorded impressions under the selected conditions. You cannot translate that into 800 prospects, 800 visits, or 800 recommendations.

Even an accurate answer with a visible source may satisfy the reader without a visit. Alternatively, a reader may later search for the business by name, telephone it, or return through another channel. Those possibilities explain why exposure and acquisition are different questions; they do not establish that either behavior occurred for your business.

Do not calculate “AI CTR” by dividing all Google organic clicks by AI impressions. Those totals describe different activity, so the resulting percentage would be misleading.

Set Up Your First Review

Start with the property covering the website or section you intend to assess. If several people manage different properties, agree on the reporting scope before exporting data. Google’s getting-started guidance recommends using Search Console to monitor how Google views and processes your website. (Google)

Open the Generative AI performance report for Search, then establish a small review record:

1. Property: Record the exact property name.
2. Period: Choose a completed comparison window, such as the latest complete 28 days against the preceding 28 days.
3. Scope: Record country, device, search-type, and page filters that are active.
4. Evidence: Preserve the chart and selected table exports with the extraction date.
5. Context: Note published content, migrations, tracking changes, or operational changes that could affect interpretation.

The 28-day window is a suggested management cadence, not a required Google setting. A monthly business review can use calendar months instead, provided the team recognizes differences in period length and seasonality.

Name the saved review clearly. “AI Search — United States — all devices — September” is more useful than “AI report latest.” Keep the same scope for the comparison period and retain the original export when you later refresh the numbers.

Read Chart and Table Totals

The report aggregates its unfiltered chart by property; page rows use page-level aggregation. A URL filter changes the chart’s aggregation. Its newest data may be preliminary. Google also applies the usual 1,000-row table limit, and exported values shown as unavailable in the interface can become zeros. (Google)

Here is a simplified illustration of why the chart and table may not reconcile by addition:

The two totals are answering different questions. One describes appearances of the property; the other describes appearances of its pages. Google’s general measurement documentation explains this distinction. (Google)

Label each number with its aggregation. Page totals can exceed the property-level chart without a reporting failure. Check that the saved table contains the rows needed for your calculation.

Preserve an annotation for unavailable values before doing spreadsheet calculations. A zero produced by an export should not automatically become evidence that a page had no exposure. If the source display was unavailable, keep that status visible in the working file.

Check the Reported URL

The page you recognize in your content calendar may not be the URL used for reporting. Google chooses a canonical URL to represent sufficiently similar versions of a page, and the selected canonical can differ from the site owner’s preference. (Google)

For example, a team may have an old service URL, a redirected replacement, and a version with tracking parameters. Before declaring that the new page has no visibility, confirm how those URLs relate and which page Google treats as representative.

This matters when a local business changes its site structure. If a Houston service page is consolidated into a broader regional page, create a before-and-after URL mapping for the review. Otherwise, a page-level decline may partly reflect where data is assigned rather than the disappearance of the underlying topic.

Treat that as a question to verify, not an assumed explanation. A targeted technical SEO review can examine the page relationships and accessibility before the content team rewrites anything.

Interpret the Local Business Context

The practical local question is whether the exposed content helps the customers you can serve.

Suppose a company wants commercial maintenance inquiries in The Woodlands. Its most visible page is a general national article about choosing equipment. That exposure might be useful, but it does not establish that the company is reaching local decision-makers ready to request service.

Review that page against three questions:

• Does it explain a problem your actual customers face?

• Does it connect readers to the relevant service and location information?

• Does the next step match the customer’s stage of decision-making?

Country data provides context, but a United States audience should not be relabeled a Houston audience. A location page appearing in the table also does not prove that its impressions came from residents of that location.

When neighborhood-level visibility matters, maintain a separately documented local search measurement process. Local SEO decisions should use evidence appropriate to location, service area, and customer intent, rather than treating this report as a map ranking tool.

A Worked Example: Exposure and Business Outcomes

The following numbers are fictional and illustrate interpretation. They are not ADM results, client results, or industry benchmarks.

Assume the same property, filters, and complete 28-day periods were used throughout.

The AI exposure measure improved. Overall organic clicks rose slightly. Qualified inquiries declined. All three can be true without contradiction.

Investigate the mismatch before assigning a cause:

• Which pages gained exposure?
• Were commercial landing pages accessible?
• Did inquiry classifications change?
• Were calls answered consistently?
• Did the business stop accepting a service?

Those are competing explanations to test. The table does not prove that AI caused fewer inquiries, nor that it generated the extra clicks.

A useful action could be to improve the connection between a well-exposed guide and the appropriate service page, then monitor what changes. Another could be to correct a booking problem discovered during the review. The evidence should decide which task deserves priority.

Our discussion of SEO metrics to measure in an AI-driven search landscape provides broader context for keeping visibility and commercial outcomes in the same conversation.

Compare Visibility With Inquiry Data

Search Console describes activity in Google’s results. Google Analytics describes what visitors do on your site. Google advises using both, while recognizing that their metrics and processing differ; clicks and sessions will not match exactly. (Google)

For a service business, add a third layer: what happened to the inquiry. Define a qualified inquiry before comparing periods. For example, it might require a genuine request for a service the business offers, inside the service territory, with usable contact information.

Keep spam, internal tests, existing-customer support, and out-of-area requests separate when they do not meet that definition. Document any changes in classification so a revised definition cannot masquerade as improved performance.

Call and form tracking can support that review, but a source label still has boundaries. “Google organic” is not equivalent to “confirmed Google AI referral.” A customer saying they used an AI tool is useful self-reported context; store it separately from automatically observed referral information.

Do not fill a missing attribution field with an optimistic assumption. A report becomes more useful when the unknown portion remains visible.

Check Data Quality

Date boundaries matter. Google documents Pacific Time for normal daily Search performance reporting. Different time zones, processing, privacy rules, and measurement methods can explain differences between tools. (Google)

Use a common reporting window for the business readout, and state any remaining mismatch. Avoid comparing a partial current day with a complete earlier day. Keep a note of the extraction date so another reviewer can reproduce your result.

Also check Google’s anomaly log when a chart changes sharply. One relevant example is the August 13–17, 2026 logging error affecting Generative AI Search impressions. Google’s August 21 update says the missing data was restored. That was a reporting issue; it should not be described as an observed loss of customers or rankings. (Google)

For a new investigation, record the affected dates and whether Google lists an issue covering them. If no matching issue is documented, keep investigating. The existence of a past reporting error is not an explanation for every later decline.

Investigate a Missing Report

Google says insufficient AI impressions can prevent the Search report from appearing. Search Labs experiments are outside its coverage. (Google)

A separate setting under Settings → Search generative AI controls inclusion. The available choices are include, exclude, or inherit the parent property’s choice. Inclusion is the default. An exclusion prevents the covered features from displaying the site’s content and links; it is distinct from AI-training controls and does not serve as a ranking or inclusion signal for other Search results, according to Google. (Google)

Confirm the intended setting with the website owner, especially where corporate and location properties share a hierarchy. Record the current choice before deciding whether a change is appropriate. An absent report should not trigger an unreviewed site-wide setting change.

If inclusion is correct, examine a few important pages for indexing and snippet eligibility. Google’s current AI guidance makes those part of eligibility and does not guarantee that meeting the requirements will produce an appearance. (Google)

The next action should follow the finding: resolve an unintended exclusion, investigate inaccessible pages, or continue monitoring limited exposure. Installing another dashboard cannot substitute for that diagnosis.

Create an Actionable Monthly Readout

Keep the recurring summary short enough that someone can act on it. Use this structure:

Scope: Property, dates, comparison window, filters, and extraction date.

Visibility: Property-level trend and the pages contributing the most meaningful changes.

Business context: Overall organic clicks, qualified inquiries, and any operational or measurement changes.

Interpretation: What the evidence supports, what remains uncertain, and the next check that could resolve it.

Actions: A small number of tasks with an owner, deadline, and expected evidence of completion.

For the fictional example above, the readout could say:

Recorded Google AI exposure increased, while qualified organic inquiries declined. We will inspect the pages with the largest exposure gains, confirm booking functionality, and review inquiry classifications. We cannot attribute the inquiry change specifically to AI from these reports.

That is an actionable finding. “AI is up 50%” leaves the owner to guess what improved and what should happen next.

Content work can then address a demonstrated need. If an exposed article leaves a service question unanswered, improve it through focused website content writing. If the issue spans identity, supporting pages, and external evidence, connect it to the broader GEO strategy.

Frequently Asked Questions

Turn the Report Into a Decision

For your next review, bring the reporting period, the pages with notable changes, and separate traffic and inquiry records. Choose one page to investigate and state the question you want to answer. Keep attribution gaps explicit, assign the next action, and agree on when to review the evidence again.

If you want a practical review of how your website is appearing in Google’s AI features and what that means for your business, contact Activate Digital Media. We can help connect the reporting questions to the website, content, and inquiry measurement work needed to answer them.

---