Skip to content
🌍Global

Europe

Google AI Overviews brings back AI images: prep your pages now

AI Search 16 November 2026 Gaurav Rajani Search Engine Roundtable
Google AI Overviews brings back AI images: prep your pages now

Nov 13, 2026: Google is testing AI-generated images inside AI Overviews again for some searchers. This was announced earlier in the summer, paused after backlash, and now appears to be returning in limited visibility.

That’s not a cute product tweak. It’s new SERP real estate that can soak up attention on queries where images matter—recipes, home improvement, product comparisons, travel ideas, “what does X look like,” and plenty of “best” and “ideas” searches.

What changed (and what didn’t)

The change: within Google Search, AI Overviews can generate images on the fly—so the results page itself can visually answer the query without needing to show (or click to) a publisher’s or brand’s page.

The non-change: this is still a test. It’s not rolled out everywhere, and the UI/placement can move around. But treat tests like weather radar: they tell you what Google wants the page to look like later.

Primary reporting here: Search Engine Roundtable on AI-generated images returning in AI Overviews.

Why business owners should care (it’s a click siphon)

If Google can show a plausible image plus a summarized answer, the user’s next step often isn’t “visit the best supplier.” It’s “refine the idea.” That’s fewer visits to your service pages, fewer category page views, and fewer product clicks—especially at the discovery stage.

We’ve already seen this pattern with AI Overviews in text form: the query gets satisfied earlier in the journey. Adding images makes that stickier. Humans don’t just skim images; we anchor on them. If the Overview shows a visual that feels “good enough,” your beautifully photographed case study is now fighting for second look.

There is an upside. Brands with strong original imagery, clean metadata, and clear entity signals can become the source Google chooses to reference, cite, or approximate. But you only get that upside if your media is machine-legible and consistently tied to the right topics.

Who wins, who loses

Likely winners: sites with unique photography/illustration (not stock), disciplined image filenames and alt text, ImageObject and Product schema where appropriate, and pages that explain the thing shown—not just “gallery” pages.

Likely losers: thin affiliate content, generic “ideas” listicles with templated images, local service businesses relying on a handful of stock photos, and ecommerce catalogs where every variant photo is near-identical and poorly labeled.

Our contrarian take: the biggest loser isn’t publishers. It’s average brand sites that assumed “image SEO” was optional. Text SEO can be decent and still survive. Bad image foundations become a hard ceiling when the SERP starts answering visually.

What everyone will miss: image trust is now brand trust

This is the part most teams won’t operationalize. When Google generates an image, it still needs grounding: real-world entities, attributes, and relationships. If your site has inconsistent product naming, missing specs, unclear licensing, or messy duplication across markets, you’re training the system (and the user) that your brand equals ambiguity.

In practice, we expect more brands to get pulled into “close enough” visuals that don’t match their actual product/service. Not maliciously—just because the model doesn’t have crisp constraints. The fix is not yelling about it on X. The fix is giving search engines structured, high-signal media assets that are hard to misinterpret.

What we’d do this month (effort vs impact)

If you’re already doing strong technical SEO, this is a 2–10 day sprint depending on site size. If your CMS media handling is messy, it can turn into a 3–6 week cleanup.

We usually recommend treating this as an “image supply chain” project: creation, naming, metadata, structured data, delivery, and measurement. The goal is simple: when Google needs a visual, your page is the cleanest candidate to borrow from—or cite next to.

  • Inventory your top 50 pages where images influence conversion (category pages, key services, comparison pages, high-intent blog posts). Make sure each has at least one truly representative image, not a decorative header.
  • Fix image semantics: filenames that describe the subject (not IMG_4837), alt text that states what’s in the image (not keywords), and captions where helpful. One good image with strong context beats five vague ones.
  • Add/validate structured data for the page type: Product (offers, price, availability), Article, Recipe, HowTo, and ImageObject where appropriate. Don’t spam schema; make it match what’s visible.
  • Harden image delivery: correct srcset, explicit width/height, modern formats where safe, stable URLs, and no blocked crawling. Performance matters because Google can’t “choose” what it can’t fetch reliably.
  • Measure the impact properly: watch Google Search Console for changes in impressions/clicks on queries that used to send traffic via “ideas” or “inspiration” searches. Tag key pages in your analytics so you can see if discovery traffic shifts to lower-funnel queries.

If your team needs help implementing this across markets, these are the tracks we run most often: US SEO retainers built around measurable traffic shifts, UK technical SEO support for SERP feature changes, and US B2B SEO programs tied to pipeline, not vanity clicks.

What this means for your site

  • Update your top money pages so the primary image is specific, original, and surrounded by plain-English explanation of what’s shown.
  • Ensure Product/Article/HowTo schema is accurate and present on the canonical URL (avoid duplicates fighting each other).
  • Stop relying on stock photography for “explainer” pages—create 10–20 reusable, ownable visuals you can deploy across the site.
  • Run an image crawl: check for blocked folders, 404 image URLs, and inconsistent CDN parameters that create endless duplicates.

Proof this isn’t just a one-off experiment

Google keeps investing in training and community education for technical SEO because these surfaces are changing quickly—AI features don’t remove technical SEO, they raise the penalty for sloppy basics. If your team is playing catch-up, Google’s own materials are increasingly structured as step-by-step curricula: Google Search Central’s learning-path playlists for Search fundamentals and Technical SEO.

Separately, publishers are already pushing for compensation when AI answers use their work, which tells you where the puck is going: the SERP becomes the destination, and everyone else negotiates for attribution or payment. Example: Yahoo Scout’s licensing approach via the Internet Licensing Collective.

If you do nothing for 90 days

You won’t “lose SEO” overnight. You’ll lose the first touch. Discovery queries that used to introduce customers to your brand will increasingly end on the results page with an AI visual and a summarized path forward.

That pushes you into a more expensive corner: you’ll rely more on branded search, remarketing, and bottom-funnel queries where CPCs are higher and competitors are already entrenched. And when leadership asks why organic “feels weaker,” you’ll be stuck explaining a UI change you didn’t prepare for.

Next step: pick 10 pages that drive revenue, and we’ll audit them for AI Overview readiness—image semantics, schema, and media performance—then give you a 2-week fix list your developer can actually ship.

Primary source: Search Engine Roundtable

Google AI OverviewsImage SEOSchemaSERP featuresOrganic traffic
Consultation with our teamTeam working together

Need any help? Contact us today

+1 (502) 780-5714

Frequently Asked Questions

FAQ: Google AI Overviews testing AI-generated images again

1. Will Google’s AI-generated images in AI Overviews reduce my organic traffic?

On image-influenced queries, it can. The results page can satisfy curiosity earlier, which often lowers clicks to “ideas” and comparison pages. Sites with strong original visuals and clear metadata can also benefit if Google uses or cites their assets.

2. What should I prioritize first to compete with AI images in Google Search?

Start with your top revenue pages: ensure the main image is accurate and original, add/validate appropriate schema (Product/Article/HowTo), and fix crawlability/performance so Google can reliably fetch and interpret your visuals.

Let’s make something great work together.

From day one to enterprise – we’re your partner in long-term tech success.

AI images return in Google AI Overviews