The Rise of AI Overviews in Search
Google's AI Overviews are changing how search results are presented, offering users quick, concise, and contextually relevant summaries of their queries.
For SEO professionals, understanding how AI Overviews work and how to optimise content for them is crucial to maintaining visibility in search results.
This article will explore AI Overviews and how they function. It will also offer practical strategies for optimising content to improve your chances of appearing in the summaries.
What Are AI Overviews?
As the name suggests, AI Overviews are AI-generated summaries that provide users with a tailored and specific response to their search queries.
AI Overviews aggregate information from multiple sources and present it in an easily digestible format at the top of the search results page, often with an option to view an expanded summary and refer to sources for the information.
The primary goal of AI Overviews is to enhance the search experience by delivering quick answers without requiring users to click through multiple web pages.
In this way, they serve a similar role to structured data and other attributes previously introduced to the SERP, like Featured Snippets, Sitelinks and Web Stories.
However, while AI Overviews may serve a similar function, there are some key differences:
Data Sources: Unlike Featured Snippets, Web Stories and Sitelinks, AI Overviews pull information from a broad range of sources, using semantic processing to identify content to cite alongside the summaries they provide. AI Overviews typically cite 7-8 web pages, though they can refer to as many as 20.
Google-generated: Whilst most other search features display content from specific websites, the text used in an AI Overview is entirely generated by Google. Google’s systems also decide which content and links appear alongside AI Overviews. As we will see, the process used to choose which websites appear in the citations is complex. A practical consequence of this is that it makes it difficult to predict.
Diversity of Content: AI Overviews can include lists, images, detailed explanations, or combinations of the three, making them more diverse.
Inside Google’s AI Overviews: How They Process and Rank Content
Google's AI Overviews use advanced machine learning models and retrieval techniques to generate summarised responses.
Specifically, they use large language models (LLMs) similar to those employed by other popular AI tools such as ChatGPT, Claude and Perplexity.ai.
The LLMs are used to understand the intent and context of a search. They then retrieve content from various sources to process and respond to the user’s search.
SEO Whizz, Rich Sanger has written several in-depth deep dives into what the process entails based on his analysis of Google’s Patent for their Search Generative Experience (the pre-cursor to AI Overviews). Check out his articles for more information.
Ultimately, the process is broken down into three key steps:
1. Query Analysis
When a user enters a query, Google's AI analyses it using large language models (LLMs) to understand the intent and context. This includes factors such as:
Query type (informational, transactional, navigational).
User's recent search history (if applicable).
Related queries and implied intent.
For example, if I type the query “Which screw anchors do I need for drywall”, Google’s AI will realise that I am asking a question (i.e., looking for information). Let’s say I am planning to hang some guitars up in my office, and earlier in the day, I also bought some guitar hangers.
Google knows this because of my search history, so it also determines that I also would like to buy some screw anchors.
2. Summary Generation
Once the context and intent of a query have been determined, Google’s AI selects a suitable LLM (one or more) that will be used to create a summary that answers the intent of the search.
The LLMs differ in terms of their training data and are tailored towards different purposes. For example, there are Informational LLMs, Creative LLMs and Image-generation LLMs.
The AI will also select the sources (articles, web pages, images or videos) which will be used to generate the response.
The sources selected for summary are chosen based on several factors that are query-dependant, query-independent and user-dependent
Query-dependant factors:
Ranking of the source
How often the source is selected in response to the query
Relevance of the result to the query
Query-independent factors:
How often the result is selected in response to other queries
Trustworthiness and authority of the source
Popularity
Freshness of the content
User-dependent factors:
The user’s recent queries
Profile information and past interactions
Familiarity with the content and sources selected
When deciding the sources to use, the AI also places value on the diversity of the content chosen and generates implied queries, which it selects based on information it has about the user and their interests.
All of this information is considered and weighed to varying degrees to generate the AI summaries you then see.
3. Why AI Overviews Don’t Always Cite the Original Source
After the summary is generated, the system begins a separate source retrieval process. This is an important point to stress. Contrary to what you would expect, the sources cited alongside the AI Overviews are retrieved after the overview is generated.
This means that the content used to create an AI Overview and the content cited by an AI Overview is not always the same.
The content cited by AI Overviews is selected based on its semantic embedding and ranking in Google’s index. Semantic embeddings indicate how similar two documents are, represented by numbers.
The less semantic ‘distance’ there is between two documents, the more similar they are. Google looks for the documents that are semantically closest to the AI Overview and also rank highly in the SERPs.
In practical terms, this means that whilst your content might be used to create a summary, the summary may not link to it if Google’s system finds another document with a higher rank or less semantic distance compared to the summary.
How to Rank in AI Overviews
Because of the complexity of the factors that go into generating AI Overviews, it is hard to predict which content will be cited by them or how they will affect your SEO.
For now, the best advice is still to focus on continuing to do good SEO. Google states that there is ‘nothing special’ for publishers to focus on that will get their content to appear more readily in AI Overviews, other than following their ‘regular guidance for appearing in search’.
However, there is more and more information for those seeking to understand how AI Overviews work. Several recent studies have been conducted to give an insight into how and how often AI overviews appear. Advanced Web Ranking has been analysing how AI Overviews have affected the results shown for 8,000 keywords since July 2024. You can view the live results of that study for free.
Overall, appearing in AI Overviews requires a strategic approach focusing on authority, relevance, and structure. Here are the key areas to consider:
1. Content Quality and Trustworthiness
Ensure your content demonstrates expertise and credibility.
Follow Google’s E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) principles.
Cite reliable sources and use factual, well-researched information.
2. Keyword Optimisation
To get into AI Overviews, focus on informational queries with mid-complexity phrases (e.g., five-word queries)
Conduct keyword research to identify terms that frequently trigger AI Overviews. AI Overviews are much more likely to be generated in response to informational or commercial queries and less likely to appear in response to navigational or transactional queries. Brand-related keywords very rarely generate AI Overviews.
3. Structured Content
Use clear headings (H2, H3) to define sections and improve readability.
Provide concise definitions at the start of your content.
Incorporate structured data markup to help search engines understand your content better.
4. Relevance and Engagement
Regularly update content to maintain freshness and relevance.
Create compelling titles and meta descriptions to encourage clicks.
Provide actionable and valuable information to increase user engagement.
5. Multimodal Content
Use images, videos, and infographics to enhance content richness.
Ensure multimedia content aligns with the topic and is optimised for search.
Adapting Your SEO Strategy for AI Overviews
SEO professionals must adapt their strategies to align with AI Overviews to stay competitive in the evolving search landscape. By focusing on high-quality content, structuring information effectively, and leveraging multimedia, businesses can increase their chances of being featured in AI Overviews and maintaining visibility in search results.
Staying informed about updates to Google's AI Overview algorithm and continuously refining SEO strategies will likely become essential to long-term success in this space.
Want to stay ahead in AI-driven Marketing? Discover how our AI solutions can optimise your SEO, PPC, and content strategies for smarter, faster results.
