MartechARTICLE

Google merges Lens and AI Mode in image search, cutting clicks to websites

Google merges Lens and AI Mode in image search, cutting clicks to websites
Image: Sabrina Santos
Article · Martech
Reading
5 min
Tables
1
In short

A combined Lens and AI Mode icon in Google's image results now opens a generative AI conversation instead of taking the user to the source page. For those who live on visual traffic, zero-click search has just gained new ground.

What was spotted

On November 19, 2026, Search Engine Roundtable, run by Barry Schwartz, reported a new behavior in Google's image search results: an icon that combines Google Lens with AI Mode. Tapping it no longer just surfaces visual search options (similar products, text in the image, pages that use that photo). It drops the user straight into a generative AI conversation, anchored to that specific image.

Damien Andell was the first to spot it, in a post on X. He described the behavior like this:

Google adds the AI MODE + LENS icon and opens an AI MODE chat with the associated image in the Google App Images tab. To quickly request AI MODE on any image in the Google App Images feed.

Damien Andell, via post on X

Schwartz tested outside the Google app and got a similar, but not identical, result: instead of the full AI Mode, the click leads to an AI Overview built from the image. In other words, the destination changes depending on the surface (app versus browser), but the pattern is the same: the image becomes raw material for a generated answer, no longer just a clickable link to whoever published it.

This isn't the Lens you already knew

Google had already placed a Lens button in image search results before, enabling reverse visual search from any photo in the grid. The difference now is where the interaction leads: the classic flow opened a layer of visual search (similar items, extracted text, source pages). The new flow opens an AI conversation that already answers, synthesizes and, potentially, removes the need to visit any website.

Classic Lens in ImagesLens + AI Mode (just spotted)
Click destinationVisual search results, with source linksAI chat (AI Mode in the app, AI Overview in the browser)
Role of the source siteCandidate for a direct clickInput for the answer, click optional
Expected interactionOne search, one clickOngoing conversation, multiple turns without leaving the SERP

Why this matters for those who live on visual traffic

E-commerce, fashion editorials, recipe portals, image banks and any brand that builds part of its acquisition on image search depend on an implicit contract: the photo shows up, the user clicks, the page gets the visit. That's exactly the link Google is shortening. If the answer about the product, the dish or the object in the photo already comes ready-made inside AI Mode, the click stops being necessary to satisfy the search intent.

This isn't an isolated phenomenon: it's the same logic as AI Overviews in text search, now extended to the image channel, historically less affected by zero-click because the answer (the image itself) always depended on opening the page to see the full context, the price or the recipe. With generative AI interpreting the image and answering in text, that dependency disappears.

In short: the visual search funnel, which used to end on the advertiser's or content creator's page, can now end inside Google's own interface, before the user even knows there's a website behind that image.

What to measure, and how

For data-driven growth and marketing teams, the practical question is: how do you know if this is eating into traffic? A few monitoring fronts make sense from now on:

  • Google Search Console, Image Search report: track the historical series of clicks and impressions segmented by search type (image) over the coming weeks, comparing against the period before this rollout.
  • Click-through rate on pages with strong imagery: product pages, recipes and galleries usually have a slice of traffic coming from Images; a disproportionate drop in that slice, without an equivalent drop in overall web search, is a sign of the effect.
  • Presence in AI Overviews and AI Mode: although there's still no native Google report for this, rank-tracking tools with SGE/AI Overviews coverage (a category that has already grown significantly in the SEO market) are starting to log responses triggered from images as well.
  • Internal before/after benchmark: since the rollout was spotted, not officially announced, teams that depend on Images should record today's numbers now, to have a clear baseline for when (and if) Google confirms the scope of the change.

What to do on the technical side (GEO applied to images)

If generative AI is going to interpret the image and answer in text, the optimization work changes target: it's not enough for the image to rank, it needs to be well described for the model that will generate the answer. That means reinforcing:

  1. Complete ImageObject and Product structured data, with contentUrl, caption, description and, where applicable, price and availability, so the model has reliable metadata and doesn't need to "guess" the context.
  2. Genuinely descriptive alt text (alt), not just for accessibility: it's one of the few text signals directly tied to the image that an AI can use when synthesizing its answer.
  3. Captions and context visible on the page itself, since AI Overviews and AI Mode tend to pull text snippets near the source image to compose the cited answer.
  4. Clear file names and entities (brand, product, variation), reducing ambiguity for Google's image classifier.

None of these practices are new: they're the same old image SEO recommendations. The difference is the reason: before, they helped the image show up in the grid; now, they decide whether the site gets cited inside an AI answer, or disappears behind it.

What's still unclear

Schwartz is direct about what he found: a behavior that was spotted, with no official confirmation from Google about its reach, the target audience of the test, or whether the feature will roll out to every market, including Brazil. His comment sums up the bigger pattern: "it's interesting to see Google continue funneling everything into AI Mode."

That sentence captures the moment well: each search surface, one at a time, gains a doorway into the AI conversation. Until Google officially confirms the rollout, the recommendation for those deciding budget and channel is to treat this as a warning sign, not a done deal: it's worth monitoring Images numbers over the coming weeks and already adjusting the structured data layer, because the cost of doing so is low and the risk of not doing it, if the trend holds, is being left invisible inside the very answer that cites the competition.

Translated from the Brazilian Portuguese original · Read the original