AI Search

Google Finds Five Ways People Actually Use Nano Banana

Google analysed 2.1 million Gemini Nano Banana conversations and found five distinct uses, led by real business work. A Google research paper studied the conversations using the Urania framework, in a privacy-preserving way, without examining the images themselves. The picture that emerges is of a tool people use for work, not just for fun.

How the study worked

The research paper analysed 2.1 million Nano Banana conversations, a sample large enough to show real patterns rather than anecdotes. It used the Urania framework, and it was conducted in a privacy-preserving way, without examining the images themselves. That last detail matters: the researchers studied the conversations and the prompts, not the pictures users made, which keeps the analysis focused on what people ask for rather than what they produce.

That method shapes what the study can and cannot say. It can describe the kinds of tasks people bring to an image model, the language they use, and the categories those tasks fall into. It cannot judge the quality of the images, or say whether the results satisfied the user. What it offers instead is a map of demand: five categories that cover how people actually use an AI image tool at scale.

The five use categories are Social and Viral, Brand and Marketing, Product Visualization and E-Commerce, Architectural and Spatial Design, and Educational and Technical Information. Each one covers a different kind of work, and together they show a tool being used across personal life, small business, and professional practice.

The social and viral category

The Social and Viral category is the one most people would expect to dominate, and the study's details complicate that expectation. Memes were the least prevalent feature in the social and viral category, which cuts against the idea that Nano Banana is mainly a meme machine. People appeared in nearly 80 percent of it, which suggests the social use is mostly about images of people: friends, family, and themselves, shared in social contexts.

That suggests the viral use of image AI is more personal than the public debate assumes. The images that travel on social platforms are often images of real people in playful or stylised form, not just captioned jokes. For marketers watching this space, the lesson is that the social audience for AI images is an audience of people sharing images of people, and content that ignores that will miss the main current.

The practical takeaway for content teams is to think about people first. If you are making social content with an image model, the use case that dominates is human-centred imagery, and the prompts that work will be the ones that handle people well: likeness, expression, setting, and style, rather than text gags alone.

Brand, marketing, and product work

The Brand and Marketing category is where the study gets most interesting for businesses. Marketing had the highest prevalence of detailed text and typography requests and precise composition control, which means marketers are not asking for vague inspiration. They are asking for specific layouts, specific wording on the image, and specific control over where things sit. That is professional work, done with a tool that many outsiders still treat as a toy.

Brand and marketing activity mostly involved small businesses, individuals, or side hustles, often fictional or unrecognized brands. That detail matters because it shows who is actually using AI image tools for marketing: not the big agencies with design teams, but the small operator who needs a flyer, a social post, or a product shot and cannot afford to commission one. For them, Nano Banana is not a novelty. It is the design department.

Product Visualization and E-Commerce is the natural partner to that marketing use. A small seller who can generate a clean product image, a lifestyle shot, or a variation on a theme has a real advantage over a seller who cannot. The study suggests that advantage is already being used at scale, by the smallest businesses first.

Design and education round out the five

Architectural and Spatial Design covers people visualising rooms, buildings, and spaces, and Educational and Technical Information covers diagrams, explanations, and instructional images. Both are practical, task-driven uses, and both show the same pattern as the marketing category: people bring a concrete job to the tool and judge it on whether the result helps them do that job.

Together, the five categories tell a coherent story. Nano Banana is used for fun, but the bulk of its serious use is work: marketing a small business, showing a product, designing a space, or explaining an idea. The meme narrative is the loudest part of the public conversation, but the study suggests it is the smallest part of the actual use.

For SEOs and link builders, the lesson is about the content that surrounds these tools. How-to guides, prompt patterns, and honest comparisons of what image models do well are all content that the study's users would find useful, because they are the people doing the work the study describes. A page that helps a small business owner make a better marketing image is a page that earns links from the communities where those owners gather.

Sources

Sources: Search Engine Journal, November 12, 2026.

Related reading on this site: Wikipedia Leads ChatGPT Citations at 10.1%, What Gets Cited by ChatGPT: Ahrefs' 1.4 Million-Prompt Study, New Sites Need Time Before Google Understands Their Value, Mueller Says, Google Releases Modern Web Guidance 1.0 for AI Coding Agents, Google's November 2026 Core Update Is Rolling Out, and Gemini 4 Carbon Arrives in AI Mode for Google AI Ultra Subscribers.

More from Linkbaits

Turn this playbook into links

Linkbaits helps you create content that earns citations and backlinks. Start with the toolkit and build your first linkable asset this week.

Start free trial