Google's New Search Console AI Reports: How to Read Them and What the Impressions Actually Mean

Short answer: On June 3, 2026, Google Search Central (authored by Hillel Maoz and Moshe Samet) launched dedicated generative-AI performance reports in Search Console. Data starts May 18, 2026. UK-first rollout. The reports show impressions only — no clicks, no CTR, no queries. That absence is confusing, and it changes how you read the number. Here's what an "AI impression" actually is, how to read the report, and what to do about the three patterns you'll see.

Where to Find the Report

Search Console → Performance → Search results → filter by Search Appearance → select AI Overviews or AI Mode. If the filters don't appear yet, your property may not be in the current rollout market. Google is expanding beyond the UK on a rolling basis.

What an "AI Impression" Actually Counts

An AI impression = your URL was included as a source card in a generated AI answer, when that answer was rendered for a user. Three things worth noting:

• The user did not have to click your source card for it to count.
• The user did not have to expand the source list — collapsed source cards still count.
• Being cited in multiple answers for the same query still counts once per rendered answer.

This makes AI impressions closer to a "brand mention" metric than a "traffic" metric. Think of it as reach inside AI answers, not clicks to your site.

Why Google Excluded Clicks and CTR (For Now)

Two probable reasons and one confirmed one:

1. Confirmed: Google says the v1 focus is helping publishers understand where they appear in AI surfaces before optimizing for click-through.
2. Likely: AI-citation click volumes are small — surfacing them prominently would invite pressure Google isn't ready to address.
3. Likely: Query-level data would reveal fan-out patterns Google considers competitively sensitive.

The Three Patterns You'll See — and What Each One Means

Pattern 1: High AI impressions, flat or falling clicks

Your content is being cited inside AI answers, and the answer is satisfying the user without them clicking through. This is the modal 2026 pattern. It's not a failure — it's a shift in what "visibility" means. Response: measure brand queries and direct traffic to see whether AI citations are lifting your entity awareness; refuse to over-index on click volume as the sole KPI.

Pattern 2: Rising AI impressions, rising branded search

The best outcome. AI answers are exposing your brand to net-new audiences, and a share of those users search you directly. Response: double down on citation signals (schema, entity, freshness) and put brand-search tracking on the dashboard beside AI impressions.

Pattern 3: Zero AI impressions on pages that rank well in Search

You rank organically but AI won't cite you. This is the diagnosis we run weekly. Almost always one of three root causes: AI-crawler blocked in robots.txt, no entity signals in Organization schema (no sameAs), or content structured as prose without extractable answer blocks. Fix in that order.

How to Use the Report Alongside the Old Search Data

• Compare AI Overview impressions to regular Search impressions for the same URL. A URL with 10× AI impressions vs Search impressions is punching above its weight — study what makes it citable.
• Segment by page. Which content types (guides, comparisons, definitions) earn the most AI impressions? That's your citable format template.
• Track over time. AI impression volume is growing 5–10% month-over-month for most sites in 2026 as coverage expands. Flat = you're losing share.
• Cross-reference with brand-search volume. Rising AI impressions + rising brand search = the flywheel is working.

What the Report Won't Tell You

• Which specific query triggered the citation.
• Whether you were the only source or one of five.
• Whether the user expanded the source list.
• How ChatGPT, Perplexity, and Claude are citing you (Google reports Google only — cross-engine visibility still requires our audit tool or a paid AI-visibility platform).

What to Do This Month

1. Confirm the reports are live on your property. If not, check back weekly.
2. Export the current 28 days of AI-impression data per URL as a baseline.
3. Cross-reference with your top organic URLs. Any large gaps get the crawler + schema + entity diagnostic.
4. Add brand-search volume to the same dashboard so you're measuring the flywheel, not just the impressions.
5. Re-audit after 45 days. This report is the first honest measurement surface we've had for AI visibility — use it.