AI Overviews

On Tuesday the 14th of May, Google announced that AI Overviews, previously known as Search Generative Experience, would be rolling out to all users in the US. You can watch the intro video here.

Based on the latest data, we’ve provided everything we know so far below, including how rank tracking tool providers are jostling to report accurately on AI Overview and how to prepare for the UK rollout (whenever that may be!).

What are AI Overviews?

AI Overviews were previously known as Search Generative Experience. It’s Google’s integration of their Gemini generative AI model into the search results. The result returns a “card” similar to a featured snippet at the top of search results. The new result aims to provide a more in depth summary on a topic from multiple sources (including websites).

AI Overviews were previously in testing via Google’s opt-in Search Labs functionality in the US before being rolled out to other non-European regions. According to Google this testing concluded that “people use Search more, and are more satisfied with their results” when AI Overviews are included.

What does it mean?

Google is likely rolling AI Overviews out because they represent a better search experience for users. Users can expect faster and more relevant answers to their questions without the need for multiple searches or poor quality results. For websites and publishers this will likely mean a reduction in organic search traffic for informational queries as summarised and more diverse answers, tailored to their query, will appear directly in the results.

However, Google claims links within AI Overviews will generate more clicks than standard SERP listings (in reality, this will be relative to your position in the AI Overview and the standard SERP). Regardless, this move reflects Google’s original goal of streamlining user searches reflected in their previous additional features like in-SERP calculators and featured snippets. 

As Google’s biggest revenue generator, “ads will continue to appear in dedicated slots throughout the page, with clear labelling to distinguish between organic and sponsored results.” It’s important that Google are able to serve at least as much inventory as they are without AI Overviews, especially as the cost of processing AI results is higher than that of a standard SERP.

What about the UK?

The UK doesn’t have a specific timeline for the release of AI Overviews, but more countries are “coming soon” and Google expects to bring them to over a billion people by the end of the year. Given that the UK has been slow to receive Labs access to AI Overviews due to concerns over GDPR, it’s possible that the UK will be later than other regions.

How prevalent will AI Overviews be?

Google has declined to quantify the number or percentage of search queries that AI Overviews will be present across, however, analysis of their presence in Labs suggests that 14-25% of queries will display an AI Overview by default with a further ~30% having an option to generate it manually. This means that up to 25% of informational traffic is at risk as a broad indicator, although this will vary by region, language and vertical and will adapt over time as more data is surfaced.

Semrush has been quick to roll out reporting on AI Overviews with initial data suggesting less than 1% of US queries trigger the AI Overview card. SEOClarity has countered that AI Overview triggering is not uniform across both logged in and logged out experiences and that rank tracking tools which use incognito (which is most of them) will be under reporting their presence. Ultimately it’s too soon to draw conclusions.

What else is coming?

AI Overviews are the only change rolling out as of today, although Google also announced other exciting search functionality coming to Labs ahead of an unspecified general release:

• Adjust your query – Soon, you’ll be able to adjust your AI Overview with options to simplify the language or break it down in more detail. This can be particularly useful if you’re new to a topic, or if you’re trying to simplify something to satisfy your child’s curiosity. 
• Multi-step reasoning – Gemini will be capable of answering more complex questions with a higher number of variables for example: “find the best yoga or pilates studios in Boston and show me details on their intro offers, and walking time from Beacon Hill.”
• Planning – You will be able to ask Gemini to create a plan, initially for meals or trips, and refine the result with follow up requests. New categories will be added such as parties, date night and workouts as the offering develops.
• Get help from AI on video – You will be able to record a video then ask a question of the video using your voice and use it as a search prompt to generate an AI Overview as well as return a list of potential websites to answer the question
• And everything else – Google says “soon, Google will do the searching, simplifying, researching, planning, brainstorming and so much more.”

Our approach at RocketMill

We have been on the forefront of preparation for AI Overviews since ChatGPT became publicly available in November 2022. Through our extensive exposure to US SERPs and search data, we are well positioned to translate the learnings from the US roll-out to the UK market. We have established a rapid feedback mechanism from the US-based brands we work with to inform UK strategies ahead of the UK roll-out.

Our approach to the AI rollout is split as follows:

1. 1. Prioritisation – by thoroughly analysing where AI Overviews are triggered across important keyword portfolios and how that impacts downstream traffic and conversions, we establish the risk and opportunity of the AI Overview and whether being present in it is valuable, achievable and represents an acceptable ROI.
   2. Content enhancement – the only way to get users to click through from the search result to your website is to provide a better experience than is available in the SERPs. This is not new but what is required to achieve it has evolved. The following elements are more important than ever:
      • Formats – we recommend formats that provide a better user experience when displaying content. This not only gives the user a reason to click through but prevents your content being entirely replicated in the AI Overview. Tables, images, videos and audio are all good examples.
      • Mark-up – search engines need to understand the context of your content in order to return it in an AI response. This requires clear and accurate mark-up for images and video alongside the traditional optimisation elements for SEO.
      • E-E-A-T – experience, the newest addition to EEAT is crucial for standing out amongst the previous ten links and is another way to demonstrate uniqueness in stance and tone of voice, establishing your brand as a destination. We work with product experts, sales teams and ambassadors to extract first-hand quotes and information to enrich search intent content.
      • Structure – technical SEO is still a vital part of a search strategy. We have seen Google reducing time spent discovering and indexing content, we optimise infrastructure to get the most value out of Google’s reduced time assessing your site.

Next steps

The rollout, available data and coverage is changing quickly. For now it’s important to assess the information currently available and make sustainable decisions based on trends rather than single data points. In the meantime, this is an opportunity to ensure that attribution is working as expected and that organic performance can be clearly tracked through to conversions by landing page.

It’s also important to note that technical infrastructure is still imperative for quick identification of content by Google. Progressing through any technical backlog will clear the decks for AI focused optimisation without the overhead of technical debt.

Keep up to date with our digestible series of AI in SEO which cuts through the overwhelming noise of AI developments to tell you just what’s important for a future-proof SEO strategy.