<img height="1" width="1" style="display:none" src="https://www.facebook.com/tr?id=2006193252832260&amp;ev=PageView&amp;noscript=1">

8 Min Read

Google Is Now Reducing AI in Search Results, According to New Research

Featured Image

Wondering how to get started with AI? Take our on-demand Piloting AI for Marketers Series.

Learn More

BrightEdge, a leading AI-powered SEO platform for enterprises, has some new exclusive research on how Google's AI Overviews—and AI in general—are impacting search.

I sat down with BrightEdge founder and executive chairman Jim Yu to learn more about how AI is shaking up search.

Which Google I/O announcements did you see as having the most immediate impact on search? How and why?

There were a lot of announcements about applications at Google I/O. However, the introduction of AI overviews has the most immediate impact on search. It represents the next chapter in search and Google's $150 billion business. 

To give you some context, if you think about it, search engines have primarily focused on using their data and knowledge graphs to provide information and links for the last couple of decades. Over time, that developed into universal results that contained visual content. However, ultimately, it was up to the searcher to do more searches to find more. 

Fast forward to today, and Google is promising to do all the searches for you, summarize the results, AND then provide an opinion. This means that much of what we know to be true in search is changing.

Its impact is already causing controversy as it fine-tunes and tests in a live environment. Hence, I believe some patience needs to be exercised while this happens. At the end of the day, AI in search is here, and it won't be going away. It's just a matter of how it settles and then where and how it will be accessible or not.

Remember that Google has been working in AI Overviews (AIO)—formally known as Search Generative Experience (SGE)—since November 2023. In parallel, this is something we have monitoring with BrightEdge Generative Parser™.

It's the industry's first and only technology that can detect patterns and changes in AI and search, and it has allowed us to give unique insights to the community in the run-up to AI overviews going live, including:

1. There was a huge spike in "No SGE" in mid-April. It then stayed high (with some variability) but recently has been consistently trending up again. Search results powered by AI decreased from 75% to 35%. 

This predicted that marketers may see less impact, and AI Overviews will only be presented for specific queries and industries while experimentation is still in motion. It also potentially indicates that Google is making way for new formats for more AI overviews in the future.

2. The amount of real estate dedicated to AI formats shrunk. This means that pixel size (amount of space on the results page) for AI Overviews was reduced significantly.

We see this trend continue as Google experiments with smaller display sizes. Their goal here is to become more effective at delivering valuable answers in concise ways. However, the relationship between AI and search will inevitably accelerate. It will get exponentially better over time.

Today, we released some new insights from BrightEdge Generative Parser™ that give more insights for marketers into what is happening and how to stay ahead. 

3. Google has gradually been reducing AI in search results from 84% to under 15%.

The likely reason is to reduce the risks of incorrect AI answers while refinements are made in a public environment. This will definitely fluctuate a lot as we monitor it daily—and it is very important to note this is snapshot and the number will go up and down day to day, week to week and industry by industry for a while.

4. AI Overview triggers and Google's industry approach to AI Overviews.

There are some common factors that trigger AI Overviews across all industries. At the same time, BrightEdge Generative Parser™ has identified which industries are seeing an impact since May 14, 2024 in the image below.

This is the first snapshot from May 30. Caveat: Healthcare and Your Money or Your Life (YMYL) vary greatly, as do Google tests and tunes.

  • 63% of keywords in healthcare show an AI Overview.
  • In B2B Tech, 32% of keywords show an AI Overview; in Ecommerce, 23%.
  • Restaurants and travel show very few AI Overviews. 

I predict variations will change, and shifts will be rapid as many outliers exist by industry

The vertical space (real estate) occupied by AI Overviews is now 20% smaller than what SGE occupied at the beginning of April. This indicates that AI is getting more precise when generating helpful experiences. This is likely because AI now caters better to people's needs, such as looking for summaries, recommendations, or conversational experiences.

Google will get better at selecting answers but work still needs to be done and AI is already predicting and answering follow-up queries before the consumers ask.

This often happens with 'what,' 'where,' and 'how' intent-based queries.

Can you share with us more about how you think Google's AI-powered search features are going to shake up SEO as we know it?

It's a searchquake, for sure. 

AI in search means that optimizing for one keyword no longer works. AI now provides multiple, predictive answers as it thinks about what else the consumer would want to know without having to do four or five additional searches. Understanding topical, conversational, and multiple types of intent-based queries is what matters going forward. 

However, we must acknowledge that getting it right will take time. Hence, the amount of AI Overviews shown will shrink and expand respectively while it finds the right fit and the most accurate answers. 

Liz Reid, Head of Google Search, addressed this issue on the Google blog last week. 

  • Google is scaling down the use of AI-generated answers in its search results,  
  • Google limited using social media posts as source material for the AI answers 
  • Google paused some answers on health-related topics and added triggering restrictions for certain queries 

AI engines are shifting to answer + experience engines, and AI has started to provide opinions on the best sources of information. Depending on your industry and where the AI overview shows up, the impact can be massive as AI results give an opinion on a brand and website. 

Where do ChatGPT/Microsoft and Perplexity fit into all of this? Are they viable search competitors?

A great question! 

Search is a $200+ billion industry, and AI has attracted new investment and entrants to the market. This has disrupted the market and, at the same time, opened up so many more choices for consumers to choose where they want to go, search for answers, and get answers. 

The rise of ChatGPT and new entrants like Perplexity means the lines between a search engine and a chatbot are blurring. However, at the end of the day, it's all search-based, so marketers need to understand that search and SEO is evolving; it's not dying, it's just changing. This is where the age-old 'adapt or die' philosophy comes in.

We released some findings last month that Perplexity referrals to brand sites were growing 40% MoM. If this continues, then it could be a serious contender and alternative to Google Search. However, remember that Google still holds a 90% market share and will still dominate. However, new options are opening up for consumers to find information in different ways, depending on the use case. For example:

  • Perplexity AI: While operating like a search engine, it focuses on informational sources and gives citations to drive website referral traffic. Branded content for authority and trust via reviews and information content is vital here.
  • ChatGPT: Conversational interaction and quick answers are key here, especially now GPT-4o extends beyond text to omni-channel. Creating content that concisely answers a wide range of questions in an engaging way is vital for marketers here. 

Regarding market share, Google still dominates at 91%, but if you're a new entrant, just taking a fraction of a percentage makes a difference. 

It's also important to remember that as the ecosystems grow, more and more partnerships emerge between 'frenemies' such as ChatGPT and Bing and Meta Llama 3 using Google's real-time data. Hence, we will see more viable options and new entrants emerge, but in certain use cases, there will be some convergence of technologies.

The big question on every marketer's mind: What's the latest on how consumer search behavior is changing due to AI-powered results and assistants? Are consumers using AI rather than clicking through to websites?


People use traditional search more than they think, whether looking for a quick answer, such as the time in London, exchange rates, or the weather. This will always happen. The same applies to consumers who just want a direct link to a website they want to see.

However, AI advancements in NLP and the development of LLMs have changed search behavior. We are seeing the emergence of new chatbot models that can offer a standalone search solution and the integration of these models into actual traditional search engines like Google and Bing.

AI allows consumers to look for information in different ways and have a choice as to what suits them best in terms of their demographics and preferences (TikTok and Meta are good examples here).  

The challenge, I think, for marketers lies in understanding that, whether it's an AI assistant or an AI-powered search engine, it's all search. It's the medium and upper funnel entry point that has changed. That's what marketers need to adapt to. 

Both AI search engines are now focusing on giving primary attention to content from credible and authoritative sources. Authority is essential for AI models, which seek out trustworthy information to provide accurate and valuable answers to user queries.

Does this stop consumers from clicking through to websites? I don't think it will prevent consumers from clicking through to websites. However, the journey will change, which will mean some reduction but many opportunities to gain referral traffic from search and AI engines.

Why? AI has become the medium that sits between a user's query and a brand's website. AI in search means consumers may have a more extended experience on a particular platform—i.e., Google and its travel planner. In other cases, like Perplexity, time on the platform may be shorter, but your brand will be linked to in its citations.

If you look at something like AI Overviews, there are still many questions to be answered on impact, as results only show for certain intent-based queries. Here, you need to monitor to see if any queries related to your brands show an AI overview. Compare traffic and conversions to what it was like before the overview. Start thinking about what content types you must include as an information-cited source. 

Across the board, citations are becoming the new form of rankings, and referrals will become a fundamental way that people will click through to websites. 

What should brands be doing about all this? Should we still be creating content, optimizing websites, and tailoring our digital presences to show up in organic results?

Absolutely yes. In fact, yes, more than ever. 

The core fundamentals of technical and website SEO are more important than ever.

Search and AI entities crave quality information. For example, Schema markup is essential so your brand can be discovered so they can see your website and understand its context relevance. The good news is that brands that have leaned into SEO and content and are already providing valuable content in your niche are off to a good start. 

In terms of optimizing for the future, quality content still matters for organic search, and authoritative content matters even more for AI and search. While each AI engine has its nuances, they open up different opportunities. Brands must be diligent and understand the various purposes and use cases they serve. 

  • Perplexity is like a research engine that gives a well-researched point of view with sources.
  • Google offers real-time answers and has a huge advantage in local search with its data.
  • Microsoft Copilot has a brilliant knowledge graph for business and information. 
  • Meta has its social graphs and app ecosystem.
  • ChatGPT has its answer engine and more and integrates with big tech platforms like Bing and more.

As a brand, you need to understand that optimizing for one search is not enough. Brands need to really hone their brand super strengths and create content that positions them as the go-to source.

To do that:

1. Ensure you become the key source of information: Make it easy for search engines and AI to find your brand.

2. Start to get predictive:  Start to anticipate what the next question your customers will have. This will be vital as AI starts offering suggestions and additional information on your brand. For example:

3. Ensure that your brand expertise is validated: Both consumers and machines need to see that your story, advice, and information are backed up with third party validation. This includes review sites, social sources, influencer and consumer reviews, and testimonials.

Brands are facing a fundamental shift where marketing teams must prepare to answer not just one or two questions but multiple types. Organizations must build and optimize content for far more formats than ever as referral traffic from AI engines becomes a big source. 

In the past, a lot of this content could be created in specific silos. Going forward, with so much content that needs to be developed, the alignment of every function must be improved, from creative, branding, design, product, and PR. There is no room left for silos in an organization that wants to win in AI and search.

As a final reminder, it is inevitable that the relationship between AI and search will accelerate. We must acknowledge that it is getting some things wrong at the moment but be aware that it is fine-tuning several things—search quality, the flow of traffic in its ecosystem, and monetization. 

AI will get exponentially better over time. We are experiencing the same growth curve with AI and search as GPUs did with Moore’s law. The number of cores in GPUs doubles every couple of years. This will happen with AI in search, but much faster, like every six months.


Related Posts

84% of Searches on Google Will Be Impacted by the Company's AI-Powered Results, Says Expert

Mike Kaput | February 6, 2024

We spoke to an expert at a leading AI-powered SEO platform to learn how Google's Search Generative Experience (SGE) will impact the future of search.

[The Marketing AI Show Episode 23]: Google Penalizes AI-Generated Content, Responsible AI Guidelines, and AI’s Impact on Local News

Cathy McPhillips | November 4, 2022

In Episode 23 of the Marketing AI Show, we talk about Google penalties for AI content, responsible AI, and AI's impact on local news.

[The AI Show Episode 92]: AI Reveals at Google Cloud Next, Google and WPP Team Up, Humane AI Pin’s Brutal Reviews & Udio, Captions Tech Spotlights

Claire Prudhomme | April 16, 2024

This week on The AI Show: Google Cloud Next showcases AI, WPP partners with Google on AI ads, and startups Humane AI Pin, Udio AI, and Captions make headlines.