Google Search Console’s Search Generative AI Report (June 3, 2026): How Detroit Businesses Finally Measure GEO and AEO Visibility

For two years, Detroit business owners have asked the same question about AI search: “How do I know if it’s working?” Until this summer, the honest answer was that you didn’t. You could watch organic clicks fall, guess that AI Overviews were eating them, and buy a third-party tool that scraped a few hundred prompts. That was it. On June 3, 2026, Google changed that by launching the Search generative AI performance reports in Google Search Console — and through late June and July, access has been rolling out to properties well beyond the original UK test group.

This is the first time Google has separated your AI Overviews and AI Mode visibility from your regular blue-link performance. For anyone doing serious Generative Engine Optimization (GEO) or Answer Engine Optimization (AEO), it is the most important measurement release of the year. It is also, frankly, incomplete — and understanding what’s missing matters as much as understanding what’s there.

What Is the Search Generative AI Performance Report in Google Search Console?

The report is a new section inside the Search Console Performance area, sitting alongside your familiar Search results and Discover reports. It counts the times your URLs appeared inside Google’s generative AI surfaces — AI Overviews and AI Mode in Search, plus generative AI features in Discover. Google splits it into two reports: one for Search and one for Discover.

Here is what you actually get:

• Impressions — how often URLs from your site appeared inside generative AI features.
• Pages — which specific URLs are being surfaced inside AI answers.
• Countries — geographic breakdown of your AI visibility.
• Devices — desktop, mobile, tablet (Search results only).
• Dates — hourly, daily, weekly, and monthly granularity, with data history beginning May 18, 2026.

What you do not get is equally important: no click data and no query-level data. You can see that your Livonia HVAC service page appeared in 4,100 AI answers last month. You cannot see what people asked to trigger it, and you cannot see how many of them clicked through. Google has said it is gathering feedback and plans to add metrics over time, but has published no timeline.

Why Is the Generative AI Report Not Showing in My Search Console?

This is the most common question we’re fielding right now, and the answer is boring: patience. The rollout is deliberately incremental. Google began with a limited test group, initially concentrated on UK properties. By June 23, 2026, SEOs in India, Switzerland, and the United States were reporting access to the new report for some — not all — of their properties. Google’s John Mueller confirmed on Bluesky that the reports are being released to sites gradually while the team reviews feedback. Barry Schwartz of Search Engine Roundtable, who covers Search Console more closely than almost anyone, didn’t gain access for his own site until July 9, 2026.

There is no setting to enable, no beta to join, and no ticket to file. If you don’t see it, your property hasn’t been included yet. The one thing you should do now is verify that your Search Console property coverage is clean — domain property verified, not just a handful of URL-prefix properties — so the data lands somewhere useful when the switch flips for you.

Note also that this report is not the same thing as the Search generative AI control that went into effect June 17, 2026 — the Settings toggle that lets you exclude your content from AI Overviews and AI Mode. That control is about participation. This report is about measurement. If you’ve been experimenting with the exclusion toggle, understand that flipping it to “Exclude” will zero out the impressions in this new report.

What Detroit Businesses Should Actually Do With AI Impression Data

Impressions without clicks feel unsatisfying to anyone raised on last-click attribution. But for local Metro Detroit businesses, AI impressions are a genuinely useful signal — arguably more useful than they are for a national e-commerce brand. Here’s why.

1. Treat AI impressions as brand impressions, not traffic

When someone in Royal Oak asks Google, “who does commercial roof repair near me and what should it cost,” and your company page is one of the three grounded citations, you got value even if nobody clicked. Your name appeared next to a trusted answer. That’s the same logic Detroit businesses have applied to billboards on I-75 for fifty years. Track the impression curve for your money pages the way you’d track share of voice.

2. Find the gap between organic rank and AI citation

The highest-leverage analysis available today is a simple side-by-side. Export your standard Search results Performance report and your new generative AI Pages report, then compare. You will find three buckets:

• Ranks well, cited often. Your content is doing its job — protect it and refresh it.
• Ranks well, never cited. This is the GEO problem to solve. The page is authoritative enough to rank but structured so poorly that the model can’t extract a quotable answer from it.
• Ranks poorly, cited anyway. Rare and interesting. Usually a page with unusually specific data — pricing, local regulations, a genuine first-hand account. Make more of those.

3. Fix the “ranks well, never cited” pages first

The pattern we see repeatedly across Michigan clients is the 1,800-word service page that opens with three paragraphs of company history before answering anything. Generative models reward extractability. Practical fixes: lead each section with a direct one-or-two-sentence answer before the elaboration; use question-shaped H2s that mirror how people speak; publish specific numbers (real price ranges, real turnaround times, real Michigan licensing requirements); add clean FAQ and LocalBusiness structured data; and cite your own sources so the model can verify you.

4. Don’t measure Google alone

Search Console tells you about Google’s AI surfaces. It tells you nothing about ChatGPT, Perplexity, Claude, or Copilot — and for B2B firms in Southfield, Troy, and Ann Arbor, those engines are increasingly where the research phase happens. Google’s report is the impression-volume backbone of a GEO measurement stack, not the whole stack. Pair it with Bing’s AI performance data and periodic manual prompt testing against the engines your buyers actually use. Run twenty real customer questions monthly and log whether you’re named. It costs nothing but an hour.

What This Means for GEO and AEO Strategy in 2026

Quick definitions, since the acronyms get muddled: AEO (Answer Engine Optimization) is optimizing to be the answer — structuring content so an engine can lift a direct response from it. GEO (Generative Engine Optimization) is the broader discipline of earning visibility and citation inside AI-generated results across engines. Neither has anything to do with geography, despite the abbreviation.

For the last eighteen months, GEO advice has been largely faith-based. Agencies made confident claims about what gets cited because no one could check. Google just handed every site owner a scoreboard — a partial one, but a real one. That changes the conversation. If your marketing partner has been billing for “AI optimization,” you can now ask them to show you the impression trend for the pages they worked on and explain the delta.

The strategic read is this: AI impressions are becoming the top-of-funnel metric that organic impressions used to be, and clicks are becoming scarcer and more valuable. Detroit businesses that adapt their reporting now — measuring citation share rather than only session counts — will make better content decisions than competitors still waiting for the click column to appear.

Your Next Three Steps

1. Log into Search Console this week and check whether the Search generative AI report has appeared for your property. If it has, export everything — the history only goes back to May 18, 2026, and you want a baseline.
2. Confirm your Settings → Search generative AI control is set to Include, unless you have a deliberate reason to opt out.
3. Pick your five highest-value service pages and rewrite the opening of each section to answer the question first. Then watch the impression line over the following six weeks.

AI search isn’t a coming disruption for Michigan businesses — it’s the current default. The difference in 2026 is that you can finally see part of the scoreboard. If you’d like help reading yours, or building a GEO and AEO program that holds up against the measurement you now have, Rosh Media works with Detroit-area businesses on exactly this.