Gemini is everywhere and on the rise.
Mid-2025, gemini.google.com took second place among AI chatbots and tools in terms of receiving traffic, and it still holds that spot today.
Gemini-driven AI Overviews (AIO) are seen by 2 billion users a month thanks to their presence at the top of Google search results, and AI Mode crossed the 75 million daily active users mark.
While these numbers may seem relatively small against ChatGPT’s 5.8 billion monthly visits, there is no way around the effect Gemini has had on Search.
Organic impressions for well-ranking B2B companies increase as they have the chance to be seen in the traditional search results, Gemini’s interface, AI Overviews, and, if users click through, even AI Mode, while click-throughs continue to drop.
Gemini is not just a tool. It’s the technology behind Google’s AI-powered features, and it’s permeated all parts of its ecosystem. For marketers, ignoring Gemini and the Search features it powers would be like ignoring Google itself.
What Is Gemini?
Released in December 2023, Gemini is a Google-owned LLM (Large Language Model) that drives multiple Search features, and its technology is also deeply integrated into Google Workspace, comes pre-installed on many Android phones, and is available as an app for iPhones and iPads.
Google is also releasing the integration of Gemini in Chrome, currently only available to users in the US on Mac and Windows.
What makes it dramatically different from stand-alone tools like Perplexity and ChatGPT is how much it has become an inherent part of Google’s infrastructure.
Just some of the things Gemini helps users with include:
- Creating and editing text, video, images, and code.
- Research, analysis, and summary of both simple and complex information, as well as large files.
- Surfacing of user-owned data across Google apps.
- Putting together strategies and processes.
- Voice-driven phone tasks like calling someone or setting an alarm.
- Building custom AI experts in the form of Gems.
Aside from being able to output different types of content (text, video, etc.), Gemini also understands multimodal input.
Gemini Models
At the time of writing, Google released Gemini 3, its most advanced model to date. It’s already available for:
- Everyone in the Gemini app
- Google AI Pro and AI Ultra subscribers in AI Mode in Search
- Developers in AI Studio, Gemini CLI, and the newly released Google Antigravity
- Enterprises in Vertex AI and Gemini Enterprise
Gemini 3 Pro is the core, generally available model of Gemini 3, while Gemini 3 Deep Think is currently only available to safety testers.
Unlike the earlier model Gemini 2.5, which had a Pro, Flash, Flash-Lite, Flash Image, and Nano version, Gemini 3 doesn’t yet have different model versions for different use cases, but Google plans to release additional models soon.
According to Google, Gemini 3 is a big step forward in terms of
- (Scientific and mathematical) problem-solving
- Multimodal processing
- Reasoning
- Coding
- Agentic capabilities (limited to Google AI Ultra subscribers)
It’s also supposed to deliver deeper and more nuanced interactions.
Disclaimer: This article was written right as Gemini 3 was released. We don’t yet know how and to what extent this release will impact Search, but we are constantly monitoring new developments.
Gemini in Search
In Search, Gemini is present as part of the technology behind
- AI Overview
- AI Mode
- The Gemini conversational Search interface, through which users can use Gemini directly
Each of these also relies on Google’s ranking algorithm, more so for AIO and AI Mode, less so for the Gemini interface.
Gemini also drives a new way of search execution dubbed “query fan-out.” Instead of performing a single search on a single query, query fan-out expands each search across multiple queries and then consolidates the answers into one. As you use it, it also tries to anticipate your next search and account for it.
Gemini in Chrome
Currently only available in the US, Gemini in Chrome uses content from your current browser tab to answer your questions. That means you can ask it questions about every webpage you visit via the Chrome browser.
It also allows you to compare or summarize information across web pages and helps you surface sites from your browser history.
From a brand’s perspective, this is important because it allows users to quickly ask Gemini whether your long-form content has the answer they’re looking for and summarize it for them.
It also helps them:
- Add reminders to their calendar based on what’s on your page (for example, a reminder to schedule a demo call).
- Understand whether your product or service meets their needs.
- Find the best solution for their problem from a listicle page.
- Compare your solution to that of your competitors across websites.
What’s more, Google plans to introduce agentic capabilities to Gemini in Chrome, enabling it to perform tasks like booking appointments or making purchases for you.
Gemini in Google Workspace
Google Workspace apps like Gmail, Docs, and Meet have built-in access to Gemini features. At the most basic level, Gemini can help you surface information and spot patterns, but it can also help you create and edit content across the apps.
Marie Haynes wrote a review of Gemini Enterprise and Gemini Business that contains some good examples of how marketers can use Gemini in Workspace to their benefit.
Gemini on Phones
On newer Android phones, Gemini comes preinstalled and can help create and edit text, images, and even code. It can summarize and look up information in Gmail and Google Drive, act as a brainstorming partner, and help you plan trips in Google Maps and Google Flights. It accepts textual, visual, and audio input, including voice search and instructions.
For iPhones and iPads, Gemini is currently only available as a conversational app. On November 5, 2025, however, Bloomberg reported that Apple is planning to close a deal with Google that would allow it to use Gemini’s technology to revamp Apple’s voice assistant Siri until its own models are sufficient.
Gemini vs the Gemini Interface, AI Mode, and AI Overview
In the rest of this article, we’ll focus on the applications of Gemini that are most relevant when it comes to getting found by prospects on the web:
- AI Overview: accessible at the top of Google’s search results page for queries that have an AIO.
- AI Mode: accessible through a tab at the top of Google’s search results page, via https://google.com/ai, and directly from the Chrome address bar (this is currently being rolled out).
- The Gemini Interface: accessible at https://gemini.google.com/app.
It’s important to keep in mind that Gemini is the technology behind these tools and that while that means there are similarities in how they operate, there are also differences. In the screenshots above, you can see that
- AI Mode and AI Overview can work with search queries, while the Gemini interface needs a conversational prompt.
- AI Mode and AI Overview are more transparent in citing sources, showcasing (some of) them in the right side panel, whereas the Gemini interface keeps them hidden behind a “sources” button.
We’ll get into more differences when we discuss how to optimize for Gemini search features.
First, let’s see how Gemini tools compare to other LLM tools and which external influences they are subject to.
Gemini Interface and AI Mode vs Other Popular LLM Tools
| Tool | LLM (Latest Model) | Owned by | Real-time browsing + known search engine(s) it pulls data from | Knowledge cut-off date | Cites sources |
| Gemini interface | Gemini 3 Pro | Yes, via Google | Jan 2025 | Yes | |
| AI Mode | Gemini 3 Pro with Deep Think soon to be released to AI Ultra subscribers | Yes, via Google | Jan 2025 | Yes | |
| ChatGPT | GPT-5 | OpenAI | Yes, via BingMight occasionally scrape Google results through third-party tools, e.g. SERPAPI | June 2024 | Yes |
| Perplexity | Uses multiple models | Perplexity AI | Yes, via PerplexityBot + Google API + Bing API | Depends on the model | Yes |
| Claude | Claude Sonnet 4.5 | Anthropic | Yes, via Brave | Jan 2025 | Yes |
| Copilot | GPT-5 | Microsoft | Yes, via Bing | Current (no cut-off date) | Yes |
Not mentioned in the table above is that Gemini, together with AppleBot, is the only big AI crawler that executes JavaScript. The crawlers of OpenAI, Anthropic, and Perplexity (and Meta and ByteDance) all don’t. (Source 1, Source 2)
External influences
| Tool | Major investors | Major publisher partnerships |
| Gemini | Funded internally by Google/Alphabet | Reddit (content/data licensing), News Corp, Associated Press (AP), Australia Associated Press (AAP), Stack Overflow |
| ChatGPT | Microsoft, Thrive Capital, SoftBank, T. Rowe Price, MGX, Dragoneer Investment Group, and others. | Associated Press (AP) (licensing), Axel Springer (POLITICO, Business Insider), Le Monde & Prisa Media, Financial Times, News Corp, AP News, Reddit (data/API) OpenAI, The Atlantic & Vox Media |
| Perplexity | Nat Friedman, IVP, Nvidia, Jeff Bezos (Bezos Expeditions Fund), Databricks, Naval Ravikant, and others. | TIME, Der Spiegel, Fortune, Entrepreneur, Texas Tribune, WordPress.com, LA Times, The Independent, Le Monde, Adweek, Mexico News Daily, CNN, Condé Nast, The Washington Post, Le Figaro, RTL Germany (NTV, Stern) |
| Claude | Amazon, Google, Fidelity, Lightspeed Venture Partners, ICONIQ. | Wiley |
| Copilot | Funded internally by Microsoft. | Axel Springer, Reuters, Hearst Magazines, USA Today Network, The Financial Times. |
The publisher deals Google has, enables Gemini to deliver up-to-date news from these sources and use their data for training purposes.
There is no explicit acknowledgement that these deals would favor the involved publishers in Gemini-driven answers, but in practice, the prominence of Reddit as a cited source in AI Overviews can’t be denied. According to Business Insider, it’s the second most-cited source in AIO, after Quora.
However, correlation doesn’t mean causation. Reddit’s prominence in both AIO and AI Mode is most likely due to a combination of:
- Google’s Gemini has easy access to Reddit’s data through the licensing deal.
- AIO’s favoring of sites with high authority and/or that offer a human experience perspective.
- AIO’s favoring of sites that already rank well in Google’s organic search results, like Reddit. (Source 1, Source 2)
Gemini’s interface will also often cite Reddit when people ask Reddit-style questions or are looking for “what people say” type of information.
The same is true for Gemini features citing StackOverflow when it comes to coding questions.
SEO for Gemini: Adapting to a New Search Experience
Gemini has directly influenced Search in multiple ways:
- Its AI Overview is part of Google’s search results page.
- Its AI Mode is integrated into Google Search.
- It’s the technology behind query fan-out, integrated into Google AI Mode and the Gemini interface.
- It’s slowly being integrated into Chrome.
What does that mean for SEOs?
Cross-Feature Influences on Search
Query fan-out
While they all have different functionalities, the fact that AI Overview, AI Mode, and the Gemini interface are all Gemini-driven makes it so that, in some cases, their impacts on Search overlap even if the features themselves differ.
The most important example of that is the reliance on query fan-out for providing answers. This technology drastically changes how search queries and prompts are answered compared to how traditional Search works, and thus how marketers should think about providing answers.
Say that someone is looking for a tool that can act as a second brain but also as a personal CRM, has an Android app, offers dark mode, and allows you to receive email alerts. A prompt like that is hard to distill into a single keyword, but with query fan-out, that’s not necessary.
Gemini will try to break this prompt down into different queries and search intents, search all the data it has access to, and compile what it finds into a single answer.
What this means for SEOs is a focus on topical authority rather than individual keyword domination. This is already a best practice of modern SEO.
- Brainstorm the core topics that might lead your prospects to you.
- For each topic, identify all possible user intents and subtopics.
- Then, do keyword research for all of those to get an idea of their popularity.
- Finally, create content for all relevant-to-your-business subtopics and intents, mapping them out and connecting them under the umbrella of their core topic.
Shopping
In mid-November 2025, Google announced the introduction of new shopping features in both AI Mode and the Gemini app (US-only for now). Users can ask Gemini a shopping-related question and get an answer tailored to their needs, including product images, prices, comparison tables, review insights, and more.
US users can now even turn to agentic AI to have Google call stores and inquire about inventory and promos, getting a summary of its research sent to them as an email or text message. You can also use it to track prices and even to buy products for you.
The Influence of AI Overviews
With AI Overview appearing at the top of the search results page, users can get instant answers to their questions without even visiting a website. As a result, the number of zero-click searches has increased, while organic clicks have dropped.
Our study of HR and Work Tech-related keywords saw an increase in AI Overview presence across the board, from BoFu over comparative queries to ToFu keywords.
A study by Kevin Indig and Erik van Buskirk + team compared clicks from search results with AI Overviews with those from search results without. While Indig admits that it’s comparing apples with oranges, the study showed a significant drop in clicks for search results where AI Overviews were present. An earlier study by Tracy McDonald came to the same conclusion.
However, Indig also found that in 80% of the cases, organic search results were still where people found the information they were looking for. This was in July 2025, and the question is if that percentage still holds today.
Lastly, AI Overview offers an entry point into AI Mode with its “Dive deeper in AI Mode” CTA at the bottom of its answer. As a result, users who would previously have turned to organic search results for more information might now click through to AI Mode to continue their search, resulting in even fewer organic clicks.
The Influence of AI Mode
At the end of Q3, 2025, Google’s AI Mode had over 75 million daily users. While some Googlers are expecting AI Mode to become the default Google search experience in two to three years (source), a report by iPullrank from July 2025 showed that 50% of the users who tried AI Mode never came back to it, and only about 9% used it more than 5 times.
That being said, Google reported week-over-week usage growth among US users, as well as a doubling of AI mode queries from Q2 to Q3 of 2025 for the US, and a study by Kevin Indig and Clickstream Solutions showed that AI Mode does manage to hold users’ attention.
Unfortunately, that usage is not resulting in traffic.
A meta-analysis by Kevin Indig found that the introduction of AI mode has significantly reduced clicks to external sites. The only exceptions are:
- Shopping-related prompts. They almost always result in clicks.
- Travel planning prompts, but these only generate clicks once the user is ready to book.
On top of that, users need fewer prompts in AI Mode than they used to need search queries to get the information they were looking for, meaning you have less chance of visibility.
The Influence of the Gemini Interface
Traffic coming directly from the Gemini interface is still very low, but that doesn’t mean it can be ignored. After all, optimizing for Gemini is optimizing for visibility across all Gemini search features.
Marketers may differentiate between AIO and Google Search, but users just see one search results page. From a user perspective, we can’t separate Gemini from Google Search anymore.
Traffic directly from the Gemini platform is still so low, but remember that optimizing for Gemini is just not optimizing for visibility in the chat interface. It’s also optimizing for visibility in Google’s AI Mode and AI Overview and for Search in general, as those features are now part of Search. While they are different things, from a user perspective, you can’t really separate Gemini from Google anymore. (Helene)
How to Optimize for Gemini
It is pretty safe to assume that what works for AI Overview will also have a positive effect on Gemini’s interface and AI Mode, but we need to keep analyzing the workings of these three features to catch potential differences. AI Mode is not just an extension of AI Overview, and the Gemini interface is not the same as AI Mode on its own webpage.
Our study on the visibility of HR and Work Tech-related queries in AI Mode showed that there is only 25% overlap between top-ranking URLs and citations in AI mode. Organically ranking for a keyword in Google doesn’t guarantee that you’ll be cited by AI Mode, and neither does visibility in AI Overview.
Kevin Indig also found that AI Overviews give more weight to regular links than to nofollow links, while the Gemini interface weighs nofollow links more than regular links, and we already saw how Google itself states that AIO and AI Mode may use different techniques leading to different results.
What we are pretty sure of is the following:
1. AI Overview and AI Mode favor high-authority sources.
Producing high-quality content, having a strong banklink profile, and building an omnipresent brand play a big role in this. Brand trust doesn’t just influence Gemini visibility, but user behavior as well.
2. AI Overview gives preference to sources that already rank well in Google.
This means that modern SEO best practices are also highly relevant for gaining visibility in Gemini search features.
3. Search intent matters more than ever.
In traditional search results, various types of pages were able to rank for the same query. Thanks to query fan-out, Gemini is able to understand the intent behind complex queries or prompts and delivers answers that match that intent.
Your chances of ranking that sales page for an informational query are lower than before.
4. Content is more likely to be featured in an AI Overview if it’s presented in a clearly structured, concise way.
This is similar to how we used to optimize for Featured Snippets.
For more feature-specific information, we’ve created dedicated articles on Google AI Mode and ranking in AI Overview.
| Optimizing for Gemini is only one part of the larger AI search landscape. Each platform operates differently in terms of crawling, citations, and ranking signals. If you’re building a broader AI visibility strategy, explore our guides on SEO for ChatGPT, SEO for Perplexity AI, and SEO for Microsoft Copilot to understand how optimization differs across ecosystems. |
Tracking Gemini Performance
Tracking Traffic
Right now, there is no way to track overall Gemini performance in one view. When people use Gemini’s chat interface, we can track those sessions through our filtering for LLMs in GA4. But if people click through from AI Overview or AI Mode, those numbers get lumped together with organic clicks in Google Search Console and can’t be tracked in GA4.
In other words, it’s currently not possible to track AI Overview and AI Mode traffic separately, and if you want to see all Gemini traffic in one view, you need to consolidate GSC and GA4 data first.
Tracking Mentions and Citations
Tools like Ziptie.dev and Peec.ai allow you to track mentions and citations of your brands across LLMs, including Gemini, with Peec being the better choice for now as it counts Gemini, AI Mode, and AIO separately, whereas Ziptie currently only offers AI Overviews. As organic clicks continue to drop and visibility has become the number one indicator of LLM search performance, you can’t rely on just tracking traffic anymore.
Aside from tracking how many times your brand gets mentioned by Gemini search features, you also want to track how it’s being mentioned. AI Mode in particular has been shown to give inaccurate replies, and you don’t want inaccuracies about your company or products to spread among prospects.
Complex Yet Unignorable
While Gemini is the technology behind AI Overview, AI Mode, and the Gemini interface, these three search features have their own ways of working and delivering results, and, as such, there is no single way to optimize for Gemini.
Gemini SEO, instead, consists of optimizing for visibility within these three systems.
While AI Mode and the Gemini interface still have a much smaller user base than, for example, ChatGPT, the mere fact that they’re owned by Google and part of Google’s search infrastructure demands our attention.
At Flow Agency, we’re following and analyzing the latest changes in the AI search landscape so we can help companies like yours increase their visibility and stay ahead of the competition. Schedule a call today to discuss your AI search needs.
