What are AI Overviews, and what do they mean for Search as we know it? Plus, the SEO team’s response.

‍

‍

A Disruptive New Search Feature: AI Overviews

During Google I/O on May 14th, 2024, a major new search feature was officially announced to be rolled out to all US-based users: AI Overviews. This new search feature is set to have implications for SEO and signifies Google's push toward the integration of AI within its existing products at large. With that, let's start from the top and delve into what exactly AI Overviews are. Following that, we will discuss some potential implications it may have on SEO.

‍

What are AI Overviews?

Simply put, AI Overviews are AI-generated overviews (or summaries) that Google is now displaying within its Search product for certain search queries. From initial observations, the queries that the AI Overview displays are mainly informational (e.g., 'my car won’t start', 'how to read as fast as possible'). 

‍

According to your particular search query, the AI Overview attempts to provide you with the best possible information. That is, in a condensed manner so that you can retrieve the information you are looking for as quickly as possible. Importantly, these AI Overviews appear prominently on the search results page at the very top. If you have used Google recently, you’ve likely already come across an AI Overview for what you were searching for. Nonetheless, pictured below is what this new search feature looks like.

‍

‍

Background

So where did AI Overviews come from? And why is Google pushing this new feature? Really, it all goes back to the public launch of ChatGPT in November of 2022. Eclipsing over a million users in only five days and steadily gaining popularity in the subsequent months, ChatGPT presented a new, revolutionary way for users to both seek out and generate information. As a result, ChatGPT emerged as a new competitor to Google. What’s worse for Google is that the creators of ChatGPT, OpenAI, received heavy investment from Microsoft, who were already a top competitor to Google. Microsoft eventually integrated the LLM (large language model) into their own search engine, Bing, which ultimately left Google with no choice but to develop their own AI model to employ within their search engine.

Google's initial move in integrating AI into Search came almost exactly a year ago with the launch of what they then dubbed SGE (Search Generative Experience) available through Search Labs. This search feature was essentially a primer to what we are now experiencing with AI Overviews. Google reports that during its testing phase, SGE fielded billions of queries and led to an overall increase in user satisfaction. Whether or not that claim was fabricated, SGE was redubbed as AI Overviews and is now available for all users in a non-experimental context.

The bottom line is that AI Overviews were born from what Google viewed as a necessity to keep its search engine relevant in this rapidly evolving era of generative AI.

‍

What Makes it Disruptive to SEO?

For SEO, AI Overviews are disruptive for various different reasons. For one, on queries where an AI Overview is displayed, it is displayed at the very top of the search result page. Likely, it is now what users are most likely to click on in the hopes of retrieving the information they are seeking out. After all, that’s all that a search engine comes down to, and AI overviews make this more efficient. To that point, the second reason AI Overviews are disruptive is the fact that - if Google’s AI is serving users a curated answer for their query and satisfying their intent in doing so - this could discourage users from actually clicking through to websites at all. Of course, such a scenario would decrease website visits across the board. Finally, AI Overviews are disruptive in that they signify the overall direction in which Google is heading. That direction being directly towards AI where - as SEOs - there is a lot of uncertainty regarding optimization methods.

From just some initial observations, however, the good news is that these AI Overviews have an affinity for displaying upon informational searches, not commercial, transactional, or navigational searches. Albeit, AI Overviews may eventually begin to display for those other three types of searches over time. For example, if I search, 'best shoes to buy online,' there is no AI Overview seeking to help me buy the best shoes for my needs. On the other hand, if I search, 'what are the best practices for grooming a dog,' an AI Overview appears. 

Essentially, for any long-tail keyword that we previously identified as a good blog post topic, an AI Overview likely appears. For searches where a user is looking to purchase something online or investigate local businesses, the AI Overview is not commonly appearing yet. However, with Project Astra (see below) having also been announced during the most recent Google I/O event, facilitating the user’s journey to purchasing something online or booking an appointment could be one of Google’s eventual goals through their integration of AI into Search.

‍

How Does it Work?

The AI Overview is powered by Google’s very own Large Language Model: Gemini. Using this AI technology to serve a conversational answer, the AI Overview pulls from the traditional results (or web pages) that already populate for the query that the user typed in. It then links to those website sources within its answer which is an important component of AI Overviews to note. Towards the beginning of the AI’s answer, it will link to these websites automatically within a collapsed dropdown. Towards the later half of the AI’s answer, however, more links will be contained but hidden underneath uncollapsed dropdowns. The image below shows an example of this using PinkTag.

‍

‍

In the example above, we can see that PinkTag is positioned second within the referenced link cards. When the link card to PinkTag is subsequently clicked on, the entire section that the AI Overview referenced is highlighted in purple.

‍

AI Overviews & Search: What’s In Store?

As it pertains to SEO and Google Search, what is ultimately in store with the arrival of AI Overviews? Quite frankly, time will tell. In the early stages of this new search feature and with many more AI enhancements on the horizon, there are a ton of unknowns, which leaves us to sort of speculate. That is, until more usage data can be uncovered. After all, it could turn out that users completely reject these AI Overviews and they gain little steam. With that, here are five potential outcomes or predictions for Google’s generative search experience.

‍

Five Predictions

1. Search Becomes More Conversational: One of the most certain predictions we’ve come to is that Google Search will become increasingly conversational. With Google AI’s ability to handle longer, more detailed queries with both a curated and conversational answer, more long-tailed, complex queries will follow. This isn’t to say that users are no longer going to be searching simple terms like ‘barbers near me’ or ‘how to ride a bicycle’, but it is to say that users may begin to supply additional details or context into their queries that they previously didn’t have an incentive to. Our content should, therefore, take into account these more complex queries.‍
2. ‍Link Positions are Sought After: As discussed in the previous section, Google’s AI Overview references some of the web pages that it used in formulating a conversational response to the searcher’s query. For SEOs, these link positions are likely going to be sought after as a result, birthing new strategies and optimization methods along the way. Still, it has yet to be revealed how effective these positions in the AI Overview are in generating clicks. Regardless, these links are primely positioned at the top of the results page which will make it almost instinctual for website owners to want to be positioned in.‍
3. Building Authority is Heightened: If users are now getting their questions or queries answered by AI, why would they even consider your site as a source of information? Furthermore, if users seek to verify the answers that AI has served them, why would they go to your website for verification? This is precisely why building authority on a given topic will become very important. If the user journey now starts with an AI answer, how can we make their next potential step in the journey to your website? The answer to that may lie in authoritativeness. Not to mention, if you are a known authority on a subject, the user may bypass AI altogether, going directly to your site to obtain the information they seek.‍
4. Featured Snippets Become a Priority: According to Google, featured snippets are not going anywhere in lieu of AI Overviews. Additionally, for those search terms that display a featured snippet—as well as an AI Overview—the featured snippet appears right beneath, which is oftentimes still above the fold. Resorting back to initial observations, we’ve found that if a website has secured featured snippet positioning, the site is also likely to be mentioned in the AI Overview. This ultimately hints at an optimization method for AI Overviews, which is discussed in the following section.‍
5. New SEO Strategies Arise: Although featured snippets are nothing new in SEO, there will be new strategies and optimization methods that arise in a search landscape that’s becoming increasingly AI-oriented. What ultimately works and what doesn’t will become a matter of trial and error. Nonetheless, within the SEO sphere, there will be lots of discussion regarding how to secure positioning within AI Overviews, to name a few. In fact, what’s recently gained steam is an acronym referred to as AIO (Artificial Intelligence Optimization). This leads us to the next section covering RedTag’s SEO team’s response to AI Overviews—including some new strategies to be experimented with.

‍

The SEO Team’s Response: AI Adaptive 

The SEO team at RedTag is no stranger to AI. Beyond using AI tools in our day-to-day workflows, we’ve seen the introduction of generative AI into Search coming from a mile away. This has effectively given us ample time to anticipate new AI search features and adjust our client strategies accordingly. Being that we are on top of such significant changes, this highlights a key differentiator of our SEO team: the fact that we are on top of and adaptive to major changes. So, in response to AI Overviews—and inevitably more AI search features—here is the SEO team’s initial plan of action.

‍

Initial Plan of Action

Monitor Search Traffic Trends

With the initial unrolling of AI Overviews, we will employ a client-wide monitoring of search traffic to see if and how this new feature impacts organic traffic. Since May 14th, 2024, is when AI Overviews became widely available in the U.S., we will specifically benchmark search traffic according to that date. Thus far, it has been reported that the websites being hit the hardest are blogging and affiliate sites. 

Identify Key Positions that Display an AI Overview

For each client, we are able to see which specific keywords generate the most traffic to their website. With that information available, we can also see if those key positions display an AI Overview. As such, if a client maintains a high-traffic position that displays an AI Overview, we will thus need to employ some of the strategies listed below in an attempt to get them an AI Overview link. Typically, the overview will mention a website that already displays in the top 10 organically, which should make this task a tad less daunting.

Employ New Optimization Strategies

Even if a link position within the AI Overview doesn’t prove effective in generating click-throughs, generative AI in Search is now a reality. Not to mention, the content we produce as SEOs competes directly with AI in this new landscape. This leads us to questions such as, 'How can we provide additional value where AI cannot?' and, 'What will make a user visit our site for information before asking an AI system?' Some of the optimization methods listed below seek to answer those questions. They also take into account this new AI landscape. 

Q&A Content

By structuring content in a Q&A format, we can increase the chances our content will be referenced by AI. Specifically, this entails using a commonly asked question as a header and including the corresponding answer as a body paragraph (see example below). Quite literally, this is the same process for optimizing for a featured snippet (or position 0) which was executed for a client below.

‍

‍

Beyond the structuring of content into a Q&A format, anticipating follow-up questions will also prove pivotal. If we’re able to answer follow-up questions within our content, the user may not need to resort to AI for retrieving the answer. Q&A content thus mimics the experience of interacting with an AI chatbot. Furthermore, for the question we are targeting with our content, we can put that keyword into an AI system and identify gaps where our own content can answer the question more thoroughly.

Information Gain 

Information Gain refers to the amount of time or effort it takes a user to 'gain' the information they’re seeking in relation to the amount of effort it takes. The concept of information gain itself partially reveals why Google has now included generative AI as a search feature. With a generative AI answer, the time it takes a user to gain the information being sought decreases significantly. They are served the answer directly without having to look through various different websites. With that, we want to serve a similar experience to users. That is, by trimming the fat off of the content we produce or formatting content in a way that a user can easily access the information they are looking for in a much quicker, easier fashion. This could even translate to adding a top-level section on all prospective content that simply summarizes the main points or takeaways.

In an article by Search Engine Land that explores this idea, they use a section from TripAdvisor as an example of quick information gain done correctly.

‍

‍

Emphasizing E-E-A-T 

Many may blindly trust the answers they are receiving from generative AI, whereas many may not. Google addresses those in the latter boat by including links to websites where its AI answer was derived from. However, to be referenced by Google or even trusted as a viable source by users (if they are seeking to verify information), we must position ourselves as actual authorities on the topic in question. This requires that we publish content according to Google’s E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) formula, doubling down on a specific area or topic where we want to become authorities. Especially with AI-generated content now running rampant, ensuring our content is reviewed for accuracy by experts (i.e., our clients) is a way to differentiate.

Crawling Optimization

Simply put, Google’s AI models still rely on crawling a site’s content to reference and formulate an answer for users. With that, we want to ensure that our clients' websites can be properly crawled with maximum efficiency. Going forward, further emphasis will be placed on ensuring client websites load fast and meet the Core Web Vitals. That is, among other aspects of Technical SEO.

Employ Tried & True Optimization Strategies

AI Overviews do not mean that we throw out all the traditional SEO methods that have been employed for years now. Rather, while we are aware of generative AI and its integration within Google, we will still apply what is proven to increase organic search traffic until those methods prove ineffective. After all, AI Overviews are not displaying on every query, and it’s still too early to determine whether this feature is here to stay for the long term. Again, time will ultimately tell the impact AI Overviews and inevitably more AI features have on search visibility, which brings us to the next part of our plan of action.

Stay Up-to-Date

Perhaps the most important part of our plan is simply staying up-to-date with any future advancements that Google announces in the near future. We anticipate more changes to come and are committed to being responsive to these changes, perhaps even proactive.

‍

Beyond AI Overviews: What Else is In Store?

As Google makes a push towards AI, this will encompass far more than just the AI Overview search feature. Rather, their LLM Gemini will be integrated into a variety of Google’s existing products such as Workspace, Google Photos, and more. However, the most significant announcements from Google I/O regarding their plans for AI in the near future are listed below.

AI-Organized Search Results

Also significant, but not discussed as thoroughly as AI Overviews, are Google’s plans to organize the search results for certain queries using AI. For search terms that activate this feature, Google’s AI will segment the existing search results into broader categories so that the user can further narrow their search—and get to the information they’re looking for faster. From Google: "This functionality will roll out for dining and recipes, followed by movies, music, books, hotels, shopping, and more later this year directly in Google Search for U.S. English results.” With search results that are organized by the AI itself, this creates unpredictability in terms of where your website will rank. However, it does reiterate that quality content adhering to the E-E-A-T formula will prevail.

‍

Gemini App

As part of the Google app, users can now access Google’s LLM Gemini in a separate tab. Again, Gemini is the main building block for all of these AI advancements that Google has and will continue to administer. Put briefly, Gemini is Google’s version of ChatGPT. With the Google application downloaded to your phone, you can now interact with Gemini directly.

Project Astra

This is Google’s plan for an AI agent that’s ‘useful in everyday life’. Quite literally, this agent would be able to process the environment akin to how our human minds do. A stunning example of this upcoming capability was seen in a demo that Google showed during their I/O event. In the video, a woman wore Google glasses integrated with the AI, and it was able to identify where she was, what objects in the room were capable of making noise, and a band name for a duo consisting of the woman’s dog and a stuffed animal. This agent technology will eventually be integrated into the Gemini app.

‍

SEO at RedTag: Stay Ahead & Visible

If not evident by this article itself, the SEO team at RedTag is committed to staying atop of the ever changing landscape that accompanies SEO. That is, with the objective of getting our clients ahead of their respective competitors in the context of organic visibility. If you are a business owner looking for a prospective partner in search engine optimization, this is our promise to you. With that, we’d love to learn more about your goals and objectives. Simply contact us by using the button below.