AI Search Optimized for AI Search

    Google Search Console Can Now Measure Multimodal Search: Why Visual Discovery Just Became a First-Class SEO Metric

    Siddhant JainSeptember 25, 20268 min read
    Last updated: October 15, 2026
    Share

    Summarize with AI

    Opens your chosen AI platform in a new tab. Sign-in may be required; article access and summaries vary by platform.

    Google Search Console Can Now Measure Multimodal Search: Why Visual Discovery Just Became a First-Class SEO Metric

    Key Insights

    • Google added a multimodal search type filter to Search Console on September 24, 2026, rolling out globally.
    • It covers Google Lens, Circle to Search, image uploads to Google Search and Chrome Search this image.
    • Data appears in the Search results performance report and the Generative AI features report, and can be exported.
    • This is a reporting change, not a new ranking factor, and more images do not guarantee visibility.
    • Brands should baseline multimodal clicks, impressions and landing pages, then test visual and content improvements.

    Search is no longer only text in and links out. Consumers increasingly search with cameras, screenshots and images, and Google has now made that behavior directly measurable for website owners.

    On September 24, 2026, Google Search Central announced a new multimodal search type filter in Search Console. For the first time, brands can isolate performance from visual and multimodal query experiences, including Google Lens, Circle to Search, image uploads to Google Search and Chrome's Search this image feature.

    The headline is simple: multimodal search has moved from an interesting behavior to a measurable search surface.

    What exactly did Google launch?

    Google added a multimodal search type filter to Search Console. The new data appears in two places: the Performance on Search results report and the report for Generative AI features.

    The filter covers four query experiences: Google Lens, Circle to Search on Android, image uploads to Google Search, and the right-click Search this image option in Chrome. As with other performance data, it can be exported for analysis in other tools.

    The integration is rolling out globally starting September 24, 2026. Sites will see these metrics if they receive traffic from these kinds of multimodal queries. Google describes the purpose as helping site owners understand how their content is surfaced in visual and multimodal search, and optimize their visual content accordingly.

    It is important to be precise about what this is. It is a reporting change. Google did not announce a new ranking system or a new ranking factor. The filter shows performance that was already happening; it does not change how that performance is earned.

    The four experiences in the filter share one trait: the search starts with something the user sees rather than something they type.

    • Google Lens. A user points a camera at an object, a label, a storefront or a product and asks Google to identify it or find related results.
    • Circle to Search on Android. A user circles or highlights something on their screen, such as an item in a video or a social post, and searches it without leaving the app.
    • Image uploads to Google Search. A user uploads a photo or screenshot directly into Search as the query.
    • Search this image in Chrome. A user right-clicks an image on a web page and searches it.

    In each case, visual context and search intent are combined. The user often cannot name the thing they want, or does not want to spend time describing it. The image does that work, and any typed refinement comes afterwards.

    Why this matters for SEO teams

    Until now, visual discovery was difficult to isolate cleanly in first-party Search Console reporting. Teams could see that image-heavy pages performed, but separating traffic that began with a camera or screenshot from traffic that began with a typed query was largely guesswork.

    That changes with this filter. Brands can now inspect clicks and impressions attributable to multimodal search directly, using Google's own data. That matters because it reveals how consumers actually discover products, locations, objects and content, rather than how teams assume they do.

    A page that looks modest in keyword reporting may turn out to be an important entry point for visual discovery. The reverse is also possible. Either way, the data replaces assumption with evidence.

    Why ecommerce and visual-first categories should care most

    Multimodal search is most natural where the product is easier to show than to describe. That includes fashion, beauty, home and interiors, auto, travel, retail, food, consumer electronics and most product-led brands.

    The behaviors are easy to picture. Someone photographs a pair of shoes they saw on a commute. Someone circles a sofa in an interiors video. Someone uploads a screenshot of a lipstick shade from a social post, or right-clicks a gadget image while reading a review in Chrome. In each case, the brand that is clearly identifiable, and whose pages clearly represent the product, has a better chance of being the one found.

    We do not yet know how large this traffic will be for most sites, and brands should not overclaim prevalence. The value of the filter is that each brand can now answer that question for itself.

    What teams should measure now

    Start with a structured baseline. Availability and granularity will depend on what Search Console exposes for your property, but a practical framework looks like this:

    • Multimodal clicks. How much traffic begins with a visual query.
    • Multimodal impressions. How often your pages appear in these experiences.
    • CTR. Whether visual appearances convert into visits.
    • Landing pages receiving multimodal traffic. Which URLs act as visual entry points.
    • Over-indexing pages. Product or category pages that attract a larger share of multimodal discovery than their overall traffic would suggest.
    • Query and intent patterns, where available. What users appear to be looking for.
    • Market and device patterns, where available. Where and how visual search is used.
    • Relationship with Generative AI features. How multimodal performance appears in the Generative AI features report alongside standard results.
    • Before and after changes. Movement following visual or content improvements, tracked over time.

    Export the data regularly so trends can be analysed alongside other channels.

    What can brands optimize?

    These are best practices to test, not guaranteed ranking levers:

    • High-quality, relevant images that clearly represent the product or service
    • Useful surrounding page context that explains what the image shows
    • Descriptive alt text where appropriate, for accessibility and understanding
    • Strong product titles and page copy
    • Consistent brand and product naming across pages and feeds
    • Accurate structured data where supported
    • Merchant Center and product feed hygiene for ecommerce
    • Crawlable image assets
    • Good image performance and page experience

    Multimodal visibility is not an image SEO hack. The whole page still matters. A strong image on a thin or confusing page is unlikely to serve the user well, and serving the user is what the system is built to reward.

    Google exposes multimodal reporting both in Search results performance and in the Generative AI features report. That means visual discovery and AI-assisted discovery are beginning to overlap operationally, at least in how brands measure them.

    We should be careful here. Google has not said that multimodal signals drive AI answers, and nobody should claim causality from the reporting structure alone. The practical point is that multimodal performance is a new measurement layer, and it belongs beside traditional and AI Search reporting rather than in a separate silo.

    What brands should not conclude

    • Google did not announce a new ranking factor.
    • More images do not automatically mean more visibility.
    • The filter is measurement, not a shortcut.
    • Traffic may be small for many sites initially.
    • Brands should compare against their own category and query behavior, not against generic benchmarks.

    The DareAISearch POV

    At DareAISearch, we think this is another step toward search becoming a multi-input, multi-surface system. Users can search with text, voice, images, screenshots, video context and conversational follow-ups. Measurement needs to evolve with that behavior.

    Search is no longer only text in, links out. Consumers increasingly search with cameras, screenshots and images, and Google is now giving brands first-party visibility into that behavior.

    We call this discipline Multimodal Search Visibility. It means understanding:

    • Where multimodal impressions and clicks are coming from
    • Which pages and products attract visual discovery
    • Which competitors dominate visual-first discovery surfaces
    • How visual discovery overlaps with traditional Search, AI Search, YouTube and video, and Maps
    • Where content, image or data gaps are reducing discoverability
    • How deployment changes affect first-party performance over time

    A practical 30-day action plan

    Week 1: Baseline. Apply the new filter, export the data and identify which pages receive multimodal traffic.

    Week 2: Cluster. Group winning and underperforming pages and inspect their visual and content patterns. Look at image quality, page context, naming and structured data.

    Week 3: Improve a controlled group. Select a set of important pages and apply improvements, leaving comparable pages unchanged.

    Week 4: Monitor. Track movement without assuming causality too quickly. Search data is noisy, and early shifts can reflect seasonality or demand rather than your changes.

    The bigger shift

    Rankings alone increasingly underrepresent discovery. A single brand can now be found through text search, AI answers, video, Maps, product surfaces and multimodal search, often within the same buying journey.

    Search intelligence therefore has to become surface-aware and input-aware. Knowing where you rank for a typed keyword is still useful, but it describes a shrinking share of how people actually find brands.

    Conclusion

    Google has turned visual discovery into something brands can measure with first-party data. The opportunity is not to chase a new trick, but to understand a behavior that was already shaping discovery.

    Brands should stop asking only "What keywords are we ranking for?" and increasingly ask "How are people searching for us, and across which input types and discovery surfaces are we visible?"

    DareAISearch helps brands measure, understand and improve visibility across traditional search, AI Search, visual discovery and emerging search surfaces. As search becomes more multimodal, conversational and fragmented, DareAISearch gives teams one intelligence and execution layer to understand where they appear, where competitors are winning and what actions can improve discoverability.

    One Unified Engine for GEO and SEO.

    Book A Demo

    Want to create your own llms.txt file?

    Use the free DareAISearch llms.txt Generator to create an upload-ready file for your website in minutes.

    Generate llms.txt

    Related Reads

    Check Your AI Search Visibility

    Find out how your brand appears across ChatGPT, Gemini, Perplexity, and Google AI Overviews.