Key Takeaways
- AI Overviews reduce clicks, especially for non-brand and informational terms
- CTR drops despite impression growth, as SERP features crowd out organic listings
- SEO tools are likely under-reporting the # of search terms and impressions for AI Overviews
- “Zero-click” is the new norm — visibility matters even without a click
- Structured content like schema, FAQs, and trusted sources fuel AI citations
- Execution gap persists — SEO success still hinges on timely implementation
No, SEO is not dying—it’s just that it has a new friend on the block. In the case of this neighborhood, the block is the SERP and the new neighbor we’re all still feeling out is the AI Overview/AI Mode. And while many companies are concerned over falling CTRs, we’re watching this evolution of the SERP—one of many we’ve seen over the years—and compiling our insights.
Where should we be concerned about organic traffic? How do we show up in AI Overviews and other tools’ answers? Should we stop doing SEO altogether?
Let's start with a state-of-the-SERP
How we got here:
- Google tried to fend off ChatGPT by releasing a half-baked version of AI Overviews that relied too much on Reddit, surfaced irrelevant information, and often offered dangerous advice on YMYL topics. The SEO and publishing community pushed back and shed light on the issue, and Google slowly reduced the percentage of AIO on search.
- Google expanded AIO during the March 2025 core algorithm rollout, but some concerns about quality remain. Google is fine-tuning its model as the community finds new bugs and issues with AIO, but now it seems the experience is much more pervasive and coherent than before—and Google might be much more aggressive in maintaining these results.
- In May 2025, Google rolled out AI Mode to the US. While it currently sits in another tab of the SERP, it is designed for multi-step queries and conversations.
“Why do we need SEO? Gen AI is going to be answering these questions.”
Also known as “My kids only use ChatGPT now.”
While AI tools are incredibly useful and are growing at an astounding rate, that does not mean SEO is dead. Or even dying, really.
Consider this: ChatGPT accounts for 0.25% of the search market, with Google accounting for 93.5%. And Google search is actually growing—searches grew 20.89% from ‘23 to ‘24, with Google Maps growing 37.63%, according to SparkToro. That’s 14 billion searches globally per day.
So, while more and more people are skipping Google altogether in favor of ChatGPT, that still leaves billions of searches every day.
That’s where Google’s AI Overviews and AI Mode are shouldering into the SERP:
- As Google’s answer to ChatGPT, AI Overviews are currently showing for about 13% of queries globally (up from earlier in the year, when the figure was around 6.5%).
- AIOs grew 116% since the March 2025 Core Update.
- The industries most affected by AIOs, according to Semrush, are the following: Science, Health, People & Society, Law & Government, and Travel.
- The prevalence of AI Overviews varies depending on the nature of the query. They’re present in about 59% of informational, top-of-funnel searches. (Which may be the types of queries your kids are entering, after all.) Informational queries and non-branded keywords are most likely to trigger AI Overviews—88.1% of AIO triggers are informational.
- This indicates a significant shift in how search results are presented and how users consume information.
Organic clicks and CTR are dropping
With AI Overviews showing in about one-third of searches, many industries are noticing organic clicks dropping. AIOs can reduce organic clicks by an average of 34.5%. Studies show that the overall number of clicks (including those from AI Overview citations) is lower than before AI Overviews were introduced.
However, since the introduction and expansion of AI Overviews, total search impressions have gone up by 49% compared to last year.
- This is due to Google surfacing more content in SERPs, often through AI-generated summaries. So while clicks are down, value isn’t lost.
- Google claims that AI Overview citations result in “higher-quality clicks” because users are gathering the information they’d normally get on your site (and competitors’ sites) in AI tools. Google’s claim is that they’re effectively pre-qualifying users, allowing your site to see more intent-driven traffic.
- Intrepid Digital has indeed seen this phenomenon of lower clicks but higher conversion rate in practice in the past weeks.
- In short, AI reduces direct clicks, but boosts brand visibility and can lead to higher CVR later.
The good news
AI Overviews and AI tools thrive on educational questions—the ones your blog posts and other top‑of‑funnel pages usually answer.
But guess what? ChatGPT and AI Overviews are pulling from exactly those sources to provide their answers. When you publish clear, authoritative content that’s technically sound and well‑optimized, you give AI engines the material they need—and a reason to credit or link back to you.
So, take heart in this fact: strong content and smart SEO still put your brand in front of searchers; the spotlight just shifts from the traditional results page to the AI overview itself.
Is LLM optimization the new SEO?
Dare we say it? GenAI optimization is very similar to good SEO. At least right now.
Good LLM marketing overlaps almost exactly with good search engine marketing. GenAI optimization is a new channel that requires synthesizing everything we know about content, SMS marketing, and SEO. You will likely already be seeing traffic coming through as referrals from Perplexity, ChatGPT, etc.
All of the basic principles that a good marketing agency or team knows apply to GenAI:
- Quality Content: You’ve probably been creating helpful, well-written content for years to rank well organically and show up in PAAs or Featured Snippets. The same quality standards apply now, more than ever.
- Unique Perspective: In order for an AI tool to surface your site or your perspective, your online content needs to be unique from other competitors and of higher quality. What facts or first-party data can you provide that others can’t?
- Technically Clean: Make sure your site’s technical SEO is clean and easy to crawl. Intensive LLM crawling uses massive resources on LLM pre-training. This means that websites are being bombarded by bots right now with additional crawling that’s not terribly efficient.
How to optimize for LLMs and AI answers
Now, about the words themselves: We do not recommend using straight AI-generated content, and definitely not at a large scale. If you choose to let AI draft your first draft of something, we suggest being heavily involved in the outlining of the page, the post-AI editing, and the brand spin only you can add. AI-generated content is a major risk because your AI tool will be pulling your brand’s supposedly unique perspective from other sources—the very definition of non-unique.
At Search Central Live in Madrid in April 2025, Google’s John Mueller confirmed, “If almost all content is auto or AI-generated, the page should be rated Lowest” by quality raters. At the same time, according to Ahrefs’ May 2025 study, 74.2% of newly crawled web pages contain AI-generated content. Only 25.8% were classified as “pure human.” This echoes patterns from 2024, where programmatic AI-heavy sites saw rapid ranking gains, then sharp crashes in visibility.
Instead, your content strategy should focus on these:
- Tap into your founders & experts, your sales staff, your marketers, and even your customers to hand-write helpful, experience-driven content. Not only will it rank better in organic listings, it will be more likely to pop up in AI answer tools.
- Lead with authority. Invest in expert endorsements, .gov/.edu links, and PR that earns immediate trust.
- Structure your content for easy skimming. Use key-fact boxes, bullet lists, and schema.
- Make sure structured data matches the visible content
Our Gen AI predictions
Visibility will continue to increase, clicks will decrease
The increase of zero-click experiences with new snippets, including AI Overviews and AI mode, will make traffic more scarce, while the increase in searches and Google’s push to create more ad inventory will increase impressions and brand touchpoints.
Search is shifting from keyword matching to intent decomposition
AI Mode’s “query fan-out” breaks complex queries into sub-intents—rewarding content that is deeply structured, modular, and semantically rich across a topic, not just optimized around a head term.
The messy middle will become a lot messier
With increased personalization and query complexity, generative search and assistants will resemble a black box. Attribution will be more challenging, journeys more difficult to map, and marketing basics will become more relevant.
Gen AI and search will be highly personalized
New developments like Google AI Mode will likely be the future of search. These assistants will be personalized with highly customized and contextual information, and this information is likely to be provided to users both proactively and reactively.
Meanwhile, more traditional search engine traffic will continue to exist, mainly in specific demographics and contexts. Information retrieval and other SEO skills will become increasingly important—as assistant-driven decision making becomes the dominant paradigm—but exactly how these skills will be practiced is still emergent.
Shift KPIs from clicks to presence
Treat GenAI presence as a new SEO KPI—not just traffic, but influence. Here are some specific tips:
- Expect & communicate an expected drop in organic traffic—and plan content & revenue models accordingly.
- Authority compounds: invest in PR, partnerships, and recognisable brand signals.
- Invest in expert-backed content, FAQ-style answers, and clear product context.
- Segregate desktop and mobile strategy. Treat desktop as a branding surface; fight for mobile if you need traffic.
- Start to monitor and optimize for GenAI visibility (LLM citations) using what tools you can.
- Know that these tools are difficult to track due to their individualized responses and the absence of standard referral paths or tracking parameters.
Google may or may not come out on top in a fragmented world of search
Google leads in transformer technology and integration of AI hardware, and it has an ecosystem that will continue adding functionality to create better customized experiences. However, the growth of ChatGPT, Copilot, Perplexity, and other search engines will create a need for marketers to prioritize multiple search engines.
How are SEO and LLM optimization different?
Let’s step back for a moment to compare the two approaches. Think of classic SEO as teaching Google to shelve your book in the right spot, every time, for every reader. LLM optimization is more like riding in a GPS‑equipped car with a smart co‑pilot who re‑routes on the fly based on your passenger’s questions (can you imagine?).
Still, SEO and LLM optimization pursue the same goal: visibility. But they get there in distinctly different ways. Traditional SEO fine‑tunes pages for Google’s predictable ranking system, while LLM optimization shapes content for dynamic, conversation‑style responses within AI tools.
This chart breaks it all down.
Conclusion
While the evolving landscape of AI in search presents a significant shift, remember that you are likely already employing many of the strategies necessary to adapt. By continuing to prioritize quality and understand the nuances of AI-driven search experiences, we can navigate this changing search environment—again.