Panel chat: AI Overviews and their impact on SEO

The rise of AI Overviews on Google is significantly altering search results and user interactions. In this video, some of our SEO experts delve into these changes, outlining the implications for organic traffic, content consumption, and provide actionable steps for marketers to effectively optimise their strategies for this AI-driven landscape.

Blog banner

The search landscape has fundamentally shifted. AI Overviews are displaying more and more frequently when you search on Google and this is naturally impacting search results and also user behaviour. To help you navigate this change, our team sat down for an in-depth discussion about the implications for SEO and what you, as a marketer, need to know.

In this video, we cover:

  • The impact of AI Overviews on organic traffic.
  • How AI is reshaping user search behaviour and content consumption.
  • What marketers should do to optimise content for AI-driven results.
  • Tips for adapting your SEO approach to stay ahead of the curve.

Watch our insightful discussion and learn how to future-proof your SEO strategy in the age of AI Overviews (full transcript is below if you prefer):

Need help with your SEO?

If you have any questions about AI Overviews or your wider search marketing strategy, feel free to get in touch on 02392 830281 or send us your details and we’ll call you.

Transcript

Chris Pitt

“We’re going to be talking about AI Overviews, their impact on search, not just specifically organic, but obviously, we’re going to lean into that quite a lot. So, let’s kick off straight away. Lee, what are AI Overviews?”

Lee Wilson

“Yeah, sure. So, they’re a search feature in Google that uses generative AI to create fast summary answers to more complex search queries. They basically function as a type of snapshot for a topic or a query. And it’s also often the beginning of a conversation where needed. Think about things like featured snippets in Google, items like ‘people also ask’, and related FAQs – this is really a progression of that into AI-generated replies. It’s an extension of what’s been happening for years. The technology is just more advanced now, facilitating helpful AI outcomes in many cases. They appear frequently when there are multiple layers to a query. For example, ‘where’s the best place to go on a family holiday in Europe with excellent bike cycle routes?’ These queries are difficult to service with a single, traditional search. People tend to pogo-stick between websites and SERPs, whereas AI Overviews are ideally placed to service those queries more effectively. Ultimately, from Google’s perspective, it’s about providing a better experience.”

Mel Soutar

“And how trustworthy is the information? As a user, when I see an AI Overview, should I trust it, or dig deeper?”

Lewis Pugsley

“In the early days, you probably wouldn’t have trusted it at all. There were hilarious examples of wrong information. It’s been around for almost two years in the US, so it’s had a lot of testing. It’s improving, but it boils down to user adoption. Someone searching in the tech industry might be more clued up, while an older, less tech-savvy user might glance past it, seeing it as an ad. Reliability depends on the query and user. But search habits are changing, so it’s one to watch.”

Lee Wilson

“The quality is definitely improving. They get billions of training set model data daily. From a learning perspective, there’s vast data access to improve and refine it. Challenges arise with new information, where data sets are fresh and skewed. Established, trusted sources improve accuracy. They appear more for ‘your money, your life’ topics, like finances and health, where accuracy matters. AI’s effectiveness depends on data set size, training model duration, and information source trust.”  

Mel Soutar

“How are they built? Is information taken from websites? If my information is in an AI Overview, does that boost my website’s performance?”

Lee Wilson

“That sounds like a Mikey question.”

Michael Carden-Edwards

“Typically, optimising for AI Overviews is a double-edged question. Don’t target them specifically. Natural AI Overview visibility comes from a well-performing website with clear data and website structure, helpful content, and demonstrated industry authority. There’s a correlation between appearing in AI Overviews and high visibility, but not necessarily causation.”

Chris Pitt

“I have a question, one I don’t have an answer to, but Lewis and I spoke a few days ago. I was testing ChatGPT search results and got irrelevant information. Lewis pointed out it’s based on my engagement. So, do different people get different AI Overviews? We know standard search is personalised. Is that happening with AI Overviews, or is there one result per search?”

Lewis Pugsley

“It’s an unknown entity. The data, in my opinion, isn’t something you can chase. If your site fundamentals are correct, you should appear. The only thing we can do is look at the data, including how often AI Overviews appear and whether your site is sourced. Third-party data is questionable, and Search Console doesn’t provide it. It’s likely personalised. LLMs, including AI Overviews and Perplexity, require qualification. Search aims for clean results, but ChatGPT depends on prompts and previous conversations. It’s opinion-based, and we’ll get more data. For now, we stand by fundamental site optimisation.”

Michael Carden-Edwards

“Benchmarking AI Overview impact is important. Third-party data is unreliable, but measures exist to understand their effect on your industry. You can export from Search Console, use Data Studio for a larger query export, and upload it into AccuRank. You can see how many queries trigger AI Overviews and your site’s position within them. This helps compare traffic before and after. AI Overviews provide quick information, raising the question of whether being featured benefits you if it reduces click-throughs. Measurement is crucial before making changes.”

Lee Wilson

“Data is more accessible. Semrush shows AI Overview opportunities. It’s similar to the featured snippets debate. We’re optimising for them, focusing on visibility. The Google pitch is that you might lose traffic but gain visibility and qualified traffic. It’s too soon to tell. We’re all optimising for it. It’s about relevance and reward. We’re seeing AI Overviews as a potential traffic loss, but also visibility gain.”

Lewis Pugsley

“It’s like the featured snippet debate. Clients asked how to get into that space. It’s about authority and core SEO principles. Click-through rates are hard to ascertain. Cited sources change frequently. It’s about brand visibility.”

Chris Pitt

“On that vanity, about metrics, do we need new metrics? Is search share more important than clicks? Does this change SEO measurement?”

Lee Wilson

“SEO has hidden behind metrics like impressions. We need to focus on business goals. AI Overviews contribute to that conversation. Prioritisation is based on client goals. There are consistent metrics like clicks and conversions, but their roles are fluid.”

Lewis Pugsley

“It comes down to client needs. AI Overviews appear more for informational queries. For some clients, this is crucial. For commercial clients, it’s less so. Informational content enriches commercial content, improving quality signals and internal linking. It’s a byproduct of overall SEO strategy.”  

Michael Carden-Edwards

“I agree. We educate clients and compromise. We aim for perfect SEO, matching competitors, but also meeting business objectives. SEO affects organic traffic, typically the largest portion. We manage client expectations, identifying relevant queries and topics. We inform clients about AI Overview impact on KPIs and traffic.”  

Mel Soutar

“Which sectors will benefit most? Travel, perhaps? Any others?”

Michael Carden-Edwards

“Finance and health are heavily impacted. Informational content is also affected. This is interesting, given Google’s emphasis on qualified sources. The Elmer’s glue example shows questionable results.”

Chris Pitt

“What’s the Elmer’s glue example?”

Michael Carden-Edwards

“The query was how to stop toppings from sliding off pizza. The answer was Elmer’s glue.”

Chris Pitt

“Did it work haha?”

Michael Carden-Edwards

“Personally, I don’t trust AI Overviews for serious topics. There are too many errors. I’m wary.”

Lee Wilson

“AI Overviews aim to summarise existing information. For medical queries, they speed up research. They compile trusted sources. However, they’re not experts. They lack human experience. There was no education on their limitations. They’re trying to speed things up, but it’s rushed.”

Michael Carden-Edwards

“Google’s advice on authority isn’t followed in AI Overview displays.”

Lee Wilson

“Exactly. AI Overviews lack Trustpilot reviews and human credibility. They present themselves as experts, leading to questionable advice.”

Lewis Pugsley

“We know the cited sources, but users might not. They might think it’s Google’s opinion. UI testing shows they’re adjusting how sources are cited. It depends on user adoption. For expert information, users might prefer trusted sources. We need more data.”

Mel Soutar

“That was the point I was about to make, actually. It’s funny because, from our perspective, it’s about how this affects us and how we build strategies for our clients. Are there additional things we ought to be doing? Is there a sub-plan to ensure people appear in these AI Overviews, if that’s right for their business? Ultimately, it depends on how users react and respond to the information being served. Like all changes to the search engine results pages, it depends on the user’s knowledge level. Are they aware they shouldn’t take the information as gospel? I think older demographics might assume it’s true because it’s at the top of the page. So, do we have any data on what percentage of searches are served in AI Overviews at the moment? Or is it too early to tell?”

Lee Wilson

“It changes all the time. I saw something indicating 70% were informational, but that’s likely based on US data versus other areas. It varies massively by industry. It’s being tested in a live environment and changes dramatically. There’s more data in the US, as it rolled out there first. One thing I wanted to pick up on, about trusting results, from a pure SEO stance, SEO enables the right people to earn the right results. There’s a symbiotic relationship between Google and the SEO community because the goal is the same: to provide relevant, useful, up-to-date, and trusted information to as many people as possible. The goals align, despite the ‘SEO is dead’ narrative. Ultimately, from an altruistic point of view, improving things is the job of an SEO person. When we look at what we do, there are two elements of SEO. One is understanding human intent and providing helpful, valuable content or results quickly. The other is optimising specific things like rich results, AI Overviews, Google Business Profiles, and other elements. But, like we said earlier, rich results and ‘people also ask’ had the same conversations we’re having about AI Overviews. We’ll likely have the same conversations in three years about some other AI-related development. It’s cyclical, but the fundamentals stay the same.”

Michael Carden-Edwards

“To touch on your point about what we do for clients, unlike other channels that are nimble and versatile for testing and pivoting strategies quickly, many of our clients are large companies with departments, budgets, and resources tied up months in advance in projects and development queues. We must exercise caution. We focus on long-term performance. You see short-term case studies on social media, but we’re investing large sums of money and resources into executing strategies. We can’t be swayed by ‘fluff.’ We can do smaller tests, like with featured snippets and ‘people also ask.’ But we need to ensure we’re doing the right thing. We know website structure, content helpfulness, and technical elements like schema and structured data are crucial. Google has been battling AI-generated and low-quality content. Our messaging emphasises that our long-term goals aren’t changing. Reporting and understanding the impact of these changes are being enhanced. Our role is to explain the ‘why’ and advise on long-term approaches. Understanding what’s happening and why is our main concern for the next 6-12 months.”

Lewis Pugsley

“Following up on that, do you think the introduction of AI Overviews, citing authoritative sites, gives us more ammunition to display a quality site when discussing AI content with clients?”

Michael Carden-Edwards

“To address Chris’s point about vanity, it depends on the client. Some want to be on the bleeding edge, while others need to be led. Overall, I think it can help a lot.”

Lee Wilson

“Relevance is key. Who’s in there now? Are they direct competitors? How relevant is it commercially? How much does it fit into immediate goals, like breaking into a new area or becoming a visible brand? Sometimes, the ranking sites are non-competitors, just established, trusted websites. We need to consider if we can displace them and if it’s worth the effort. It’s about relevance and achieving meaningful results.”

Mel Soutar

“Going back to Mikey’s point, are we saying there’s no shift in who occupies the top of Google? Are they the same authoritative sites Google already favoured? Has it not shifted the results page as much as we thought?”

Michael Carden-Edwards

“It depends. Every industry and country behaves differently. There are basic SEO building blocks, but success varies. Some industries allow for riskier strategies, while others require squeaky-clean approaches. I’ve seen varying impacts on my clients. Measurement is crucial. Benchmark and measure what’s happening in your industry and proceed based on that information.”

Mel Soutar

“It feels like we’re at a ‘business as usual’ position. Well-optimised sites with helpful content will continue to gain visibility. The question is visibility versus visits. Only time and more data will tell.”

Lee Wilson

“It comes back to algorithm chasing versus understanding your audience and providing unique, credible, and valuable content. Creating useful tools that serve your audience will perform regardless of AI Overviews or rich results. It always comes back to value. Algorithm chasing is for quick wins. We saw this with AI content, but it quickly became noise. It has to be based on understanding and value, or it’s short-term and potentially damaging.”

Lewis Pugsley

“To add to that, the only change is that some webmasters are blocking LLMs. We need to be aware of what these bots can crawl. But otherwise, on-page, off-page, and technical SEO remain the same.”

Mel Soutar

“Excellent. Thank you.”

Chris Pitt

“To wrap up, LinkedIn is full of AI hype. How hard should we push AI to our clients?”

Lewis Pugsley

“Use it to back up your fundamental goals. Rely on data and performance, which will hopefully come.”

Chris Pitt

“Lee?”

Lee Wilson

“It comes back to justification and your goals. Leverage data if it’s significant, otherwise adapt your strategy.”

Chris Pitt

“Mikey?”

Michael Carden-Edward

“Awareness, reporting, and understanding are key.”

Chris Pitt

“Thank you. This was a really interesting conversation.”

Michelle Hill profile picture
Michelle Hill

Michelle joined Vertical Leap in 2011 as Marketing Manager, having spent the previous 15 years of her marketing career in the recruitment, leisure and printing industries. Her passions include dogs, yoga, walking, cycling, the beach, mountains and tapas.

More articles by Michelle
Related articles
Angry looking robot with staring eyes

8 ways marketers can compete with AI-written content

By Lee Wilson
optimise for ai tools blog header

How to rank in AI Overviews, ChatGPT and other AI tools

By Dave Colgate
Woman on phone with lots of different screenshots flying out

Search everywhere: The ‘new SEO’ and optimising for multichannel visibility

By Chris Pitt