Google AI Overviews appeared on 12.95% of United States searches. The figure comes from a 2026 measurement by analytics publisher shattered.io. The share sits just above one in eight US queries.

What does the dataset measure?

The 12.95% reading comes from shattered.io's proprietary tracking of Google results pages. The methodology does not appear in the published material. Sample size, query universe, and measurement window all remain unspecified.

Shattered.io did not state whether the figure includes personalized sessions, incognito sessions, or both. The release also lacks a device-level split between mobile and desktop.

Without disclosed parameters, the 12.95% share functions as a directional signal rather than a defensible industry benchmark, leaving the dataset un-auditable from public information alone.

Read 12.95% as one vendor's snapshot rather than industry consensus.

Which verticals absorb the share?

The shattered.io release contains no breakdown by industry, query length, or search intent. Publishers cannot determine from the source which categories receive AI Overviews at the highest rate.

The omission leaves affiliate sites, news publishers, and e-commerce operators without a segmented risk profile. Preparation depends on knowing where the feature concentrates its appearances across the search results.

Vertical-specific data will determine which publishers face the largest traffic displacement. The 12.95% headline does not separate informational from commercial queries, so caveat the figure before acting on it in any planning cycle.

Segment-level risk remains unmapped by this report.

What does the figure not include?

The supplied material omits absolute query volume across the US search market. No metric on user clicks, scrolls, or follow-up refinements accompanies the share estimate.

These omissions carry more weight than the headline number for commercial planning. A 12.95% appearance rate does not equal a 12.95% traffic loss, since citations remain embedded above source links in many query formats on desktop and mobile.

Treat the figure as a floor estimate rather than a definitive industry reading. The cautious operator waits for click-through displacement studies across ten-thousand-query samples before drawing revenue conclusions.

Claim the share only when the methodology is in hand.

How does the source compare to prior readings?

The shattered.io release does not reference earlier AI Overviews share measurements from any third party. Independent cross-references do not appear in the supplied material.

Comparisons to prior vendor readings cannot be drawn from this source alone. Each measurement firm uses distinct detection rules, query pools, and refresh cadences that shape its share estimate.

The 12.95% reading therefore stands alone within the public record supplied here. Independent replication will require shattered.io to publish its full data collection process, including detection thresholds, query filters, and refresh intervals.

Until replication arrives with disclosed query pools and detection logic, the industry record holds no comparable figure against which to test the 12.95% reading.

One vendor's data does not constitute a market benchmark.

What to monitor next

Track four signals over the coming weeks. First, watch for full methodology disclosure from shattered.io within 30 days. Second, watch for replication by an independent measurement firm. Third, watch for any Google statement addressing share figures across regions. Fourth, watch for click-through displacement studies that cover at least ten thousand queries per segment.

A disclosed query universe will determine whether 12.95% anchors future industry debate. Until shattered.io publishes its methodology, the number occupies the record as a single-vendor estimate demanding independent verification.