If images on your site drive discovery, you no longer need to guess. Google Search Console now surfaces a dedicated “multimodal” search-type filter that isolates traffic originating from image-based queries — Lens, Circle to Search on Android, image uploads to the Search app and Chrome’s right‑click “Search this image.” The change began rolling out globally on September 24, 2026 and appears in the Performance report and the Generative AI features report.

What changed — and what data is included

Search Console separates multimodal image queries from traditional text queries. When users start searches with photos or image uploads, those sessions are now tagged and available via the new multimodal filter in Performance and in the Generative AI features report. The dataset covers interactions from Google Lens, Android’s Circle to Search, image uploads to Google Search and Chrome’s “Search this image.” You can export the filtered data for offline analysis.

Why marketers and publishers should care

Visual queries reflect different intent and discovery paths than typed searches. People using images are often seeking identification, visually similar products, or quick visual answers — behaviors that can produce distinct engagement and conversion patterns. Until now, teams had to infer image-driven discovery from indirect signals like image pageviews or traffic shifts. The new filter gives a direct signal you can segment alongside your other performance metrics.

Practically, that means you can answer questions you couldn’t measure reliably before: which pages are found via images, which specific images surface in visual results, and whether multimodal visits convert or engage differently than text-driven traffic. Google hasn’t added new multimodal-specific conversion metrics in this announcement; however, because the filter applies to existing Performance metrics you can segment impressions, clicks and CTR by query type and combine the export with on‑site analytics for downstream behavior analysis.

How to act now

  • Apply the multimodal filter in Search Console: Open the Performance or Generative AI features report, select the new multimodal search type filter and export the data. Use that export to compare impressions, clicks and CTR against text-query traffic in your analytics workspace.
  • Treat images as discovery assets: Optimize filenames, alt text and captions; use high-quality images that load quickly on mobile. These steps support accessibility and improve the odds of images being surfaced in visual search.
  • Map images to intent: Identify pages where visuals carry intent — product pages, recipes, how‑to steps, galleries — and segment them in the multimodal report to see where image searches contribute most.
  • Join multimodal exports with site analytics: Combine exported Search Console data with time on page, bounce, conversion and funnel metrics to measure downstream value and identify measurement gaps.
  • Monitor overlap with generative features: Because the filter is also available in the Generative AI features report, watch how visual answers pair with generative snippets in your verticals and whether that affects click behavior.

Make multimodal data part of routine reporting and experimentation rather than a one‑off check. Where multimodal traffic is material, test image variants and page treatments to see if discovery and conversion move together.

What to watch next

Google’s announcement invites feedback and signals the dataset may evolve. Expect additional documentation and possible refinements as the company collects usage signals from site owners. For now, prioritize detecting where multimodal queries already matter, fold those insights into SEO roadmaps and analytics pipelines, and plan image-specific experiments where the data shows material impact.

Start by exporting a baseline from the new multimodal filter, compare it to your text-query metrics for the last 30–90 days, and add a visualization of multimodal share to your regular performance dashboard. That will turn a new reporting capability into repeatable insight.