How Visual Assets Shape AI Search and the Ads Inside It
Google Lens handles over 25 billion visual searches a month, and AI search ads now carry images. How visual assets shape AI search and its ads.
Google began moving campaign-level broad match and Automatically Created Assets into AI Max in September 2026, and Dynamic Search Ads follow in February 2027, according to Google's own announcement. The words in a search ad are being handed to automation. The pictures are heading the same way. In March 2026, Google's global advertising vice president Dan Taylor said Lens handles more than 25 billion visual searches a month with 20% of them showing commercial intent. On 24 September 2026, Google began reporting those searches in Search Console. AI systems now read the images on a page, match them to people's questions, and increasingly show ads right next to them.
How AI Search Reads an Image
Since September 2025, AI Mode has used what Google calls a "visual search fan-out", which recognises subtle details and secondary objects in a photo as well as its main subject and runs several searches in the background. A person can start an AI Mode search by uploading an image or taking a photo, and refine the results in conversation.
What the system finds depends on how clearly the image is described around it. When a photo, the words around it, its file details, and the page's structured data all point to the same product, place, or service, an AI system is much less likely to get it wrong. Google's own guidance from May 2025 points the same way. It advises site owners to support their text with high-quality images and videos, keep Merchant Center and Business Profile information up to date, and make sure structured data matches what is visible on the page.
What Google Now Measures
On 24 September 2026, the Search Console stated it added a multimodal search filter to both the standard performance report, and the generative AI features report. It covers searches through Lens, Circle to Search on Android, image uploads to Google Search, and the right-click "Search this image" option in Chrome, and it is rolling out globally. It builds on the generative AI report this blog examined in the piece on what Google's AI search report shows. For the first time, a team can see which of its pages people reach by pointing a camera at something or uploading a picture.
Why AI Search Ads are Turning Visual
On the advertising side, images are becoming the unit of competition. Google's Shopping Graph holds more than 50 billion product listings with more than 2 billion refreshed every hour, and those listings supply the products AI Mode shows when someone shops in conversation. Since a September 2025 update, Google's Asset Studio lets advertisers generate lifestyle images from an uploaded product photo, transform up to 100 images at once, and use on-brand images as a style reference for new ones.
As Search Engine Journal reported on 1 July 2026, ChatGPT's ads already pair a headline and a short description with an image and a link. This blog covered how those ads are placed in the piece on ChatGPT Ads and paid placement in AI search. OpenAI job postings describe text, image, video, native, conversational, and interactive ad formats, and OpenAI says it will "evolve the program over time to support additional formats". Microsoft began piloting Showroom ads in Copilot in April 2025. They open a rich, sponsored product experience once a person shows purchase intent. Not every challenger is following. Perplexity is staying ad-free and pursuing subscriptions and enterprise deals instead, according to Similarweb analysis reported by Search Engine Journal in September 2026.
What this Means for Creative Teams
Four practices follow from this. First, favour original, specific imagery over generic stock, so both a customer and a machine can tell exactly what the product, place, or service is. Second, give each image one clear subject, and keep its filename, alt text, caption, and structured data describing that same subject. Third, keep product feeds and business listings current, since AI search draws on them directly. Fourth, record which images are photographs and which were generated or edited with AI, now that ad platforms can generate imagery on request. Provenance standards such as Google's SynthID and C2PA Content Credentials are built to carry that record with the image itself. For a fintech or web3 brand, an AI-generated image of an app screen or a card can show features that do not exist. That makes tracking visual asset records a strict compliance requirement, and a creative discipline as well.
What to Watch Next
Three developments will show how quickly this moves. The first is whether OpenAI's larger-image and conversational ad formats move from job postings and mockups into testing, which it has not yet dated. The second is what Search Console's multimodal data reveals over the coming months about how many visits start with a camera rather than a keyboard. The third is the Dynamic Search Ads migration in February 2027, which moves one more legacy campaign type into AI Max, and leaves creative quality as one of the clearest levers an advertiser still controls.
FAQs
How many visual searches does Google Lens handle? More than 25 billion a month with 20% showing commercial intent, according to Google's global advertising vice president Dan Taylor in March 2026.
Can I see visual search traffic in Search Console? Yes. Since 24 September 2026, a multimodal search filter in the performance reports shows traffic from Lens, Circle to Search, image uploads to Google Search, and Chrome's "Search this image" option.
Do ChatGPT ads use images? Yes. The current ad unit pairs a headline and a short description with an image and a link, and OpenAI job postings describe image, video, and conversational formats in development.
What makes an image easier for AI search to use? Original, specific imagery with one clear subject. Its filename, alt text, caption, and structured data should all describe that subject, and the product feeds and business listings behind it should be kept up to date.
Not sure how to start planning and executing your on-brand visual assets to meet these specifications? Reach out at hello@marcomfin.tech.
This article was first published on MarcomFintech at https://marcomfin.tech/blog/visual-assets-ai-search-ads. © MarcomFintech. All Rights Reserved.


