The traditional way of performance marketing through search engines is fundamentally changing with the introduction of AI, and businesses need to conduct search engine marketing strategies to remain competitive. This whitepaper discusses, in detail, the following:
Table of Contents
- Changes in user behavior and technology
- Impact on marketing
- SEO
- SEM
- What to do next?
- LLM Readiness Monitoring
- Additional tools
- Resources and further reading
Executive Summary
- Research suggests that AI-generated search results and chatbots are changing how people search, moving from simple keyword searches to complex queries with direct answers.
- It seems likely that users are clicking less on ads and organic results, as AI tools like ChatGPT and Google AI Overviews provide immediate answers.
- The evidence indicates marketers need to adapt SEO and SEM strategies to focus on user-centric, high-quality content for AI integration.
TL;DR
AI-driven search is reshaping how people find and interact with information, moving from simple keyword entries to rich, conversational queries that deliver direct answers via chatbots and AI summaries. As a result, traditional click-through rates on ads and organic listings are declining, compelling marketers to pivot towards user-centric, high-quality content designed for AI consumption.
Key insights:
- Evolving Search Behavior: Users increasingly rely on AI tools (e.g., ChatGPT, Gemini) for complex, multi-part queries, with AI prompts averaging 23 words versus just 4.2 for standard searches. While Google and Bing remain dominant for general information, AI platforms excel at specific tasks like product comparisons and learning
- SEO Adaptation: To win AI visibility, content must be comprehensive, authoritative, and naturally phrased. Strategies include entity-focused writing, up-to-date citations, engaging multimedia (videos, infographics), structured data (HowTo, FAQ schemas), and user interaction features (comments, polls) to boost AI algorithms’ confidence.
- SEM Evolution: AI automation in Google Ads (Performance Max, Demand Gen) is driving up impressions but lowering CTRs. Advertisers should integrate first-party data, maintain tight keyword-to-ad relevance, isolate high-performing terms into dedicated campaigns, and supply diverse creative assets for AI optimization.
- Actionable Next Steps: Continuously audit LLM readiness using tools like SEMrush or, even better, Prebo Digital’s audit platform, implement structured data via Google’s markup helper, and track AI referrals in GA4 to measure real-world impact.
AI Results: An Introduction
By now, we’re all aware of the significant impact that AI technology is having on our daily lives, both in the workplace and outside of it. I, for example, recently used Gemini Live to ask how to best prune my growing potted lemon tree so that it would get into the right shape without becoming heavy or lopsided.
Besides assisting with my gardening, AI tech significantly impacts how people look for, find, and process information to make decisions. In my world of marketing and digital, this directly relates to how my clients spend their marketing budgets and how they generate an optimal return.
Based on my observations over the past weeks and months, I realized the need to gain a much deeper understanding of not only the shifts in user behavior but also what this means for my business, my clients, and the wider industry.
The below is an attempt to establish what’s changing in the content, paid search, and SEO landscape at this point in time and what we need to do to ensure our clients and our own platforms gain optimal inclusion in AI platforms and reduce reliance on traditional search engine marketing strategies.
Before diving into the technical impact and recommendations, we need to establish what is happening.
Sidenote: this article was written (obviously) with assistance from Grok3 deep research and ChatGPT 4oPlus and I’ve tried to apply whatever I’ve learned into this article as examples to make sure this article is eligible for LLMs 😉
Summarize this article in ChatGPT
1. Understanding the Shift in Search Behavior
AI technologies, such as conversational chatbots (e.g., ChatGPT, Gemini) and AI search engines (SearchGPT, Perplexity, etc) are transforming how users find information.
Traditionally, searches involved entering keywords and clicking on one or two website links, but now, users often receive comprehensive answers directly within chat windows or AI-generated summaries at the top of search results. This shift is driven by AI’s ability to handle complex, multi-part queries, reducing the need for users to navigate multiple websites.
Looking at recent research, the survey in “[AI Search Gaining Traction, Not Replacing Google” (https://searchengineland.com/ai-search-gaining-traction-not-replacing) found:
- 71.5% of respondents use AI tools for search, with 14% daily, but 79.8% prefer Google or Bing for general information. This indicates AI is not replacing traditional search but is significant for specific queries like shopping comparisons, where tools like ChatGPT and Claude are preferred. This partial shift supports the hypothesis, showing varied user behavior by query type.
- The shift is evident in query complexity, with ChatGPT prompts averaging 23 words versus 4.2 for searches, and 70% of prompts being unique, as per Semrush analysis in the ChatGPT article.
- Generational differences show Gen Z (82%) and Millennials using AI more for professional/educational queries, while Baby Boomers prefer traditional search, per the AI search survey. This suggests a nuanced impact, with AI excelling in specific areas like shopping comparisons and learning, reducing clicks in those contexts.
For businesses, this has a significant impact as well, which is already visible in the way businesses are starting to receive increasing incoming traffic from AI Chat platforms.
Large omnichannel retailer with many 1000s of product pages.
What’s most interesting for the above retailer is that the majority of referral traffic from ChatGPT is for more complex products like buying furniture online, and the conversion rate is 45% higher compared to normal SEO traffic (remember that includes branded traffic, too!).
On the contrary, below vehicle service business with minimal content on the site sees very little traffic and results from LLMs and AI conversational platforms, yet:
“Although still very behind Google, it’s impossible not to see how ChatGPT adoption as an assistant for information discovery, rather than just tasks execution; although with a more sophisticated, granular search behavior that we need to start tackling better: besides ensuring our content crawlability, optimizing our brand citations, better structured content that facilitates their findability.”
Aleyda Solis, International SEO Consultant & Founder of Orainti
Source
(Keep reading to see why I included the above quote)
Interestingly, in the data we’re able to measure in GA4 for various businesses we work with, is that the users coming through from platforms such as ChatGPT are much more skewed towards desktop than mobile. Where an average website gets about 75% of its traffic from mobile, looking at ChatGPT referral traffic, that turns to a 50/50 split.
Although I doubt anyone would really dispute the rapid growth in people utilizing AI platforms for work, shopping, personal life decisions, and more, it’s good to be able to paint a picture in May 2025 on where we stand with adoption so we know how to use this in marketing strategies.
2. Impact on (Search) Marketing
Having established the growing impact of these platforms on how people find answers, this begs the question of how it will impact digital marketing strategy and marketers focus to not miss this opportunity and trend.
Because this is a wide area to investigate that can quickly get relatively technical and detailed, I will be breaking it down into the impact this and its wider principles has on Search Engine Optimization (Organic search traffic) and on Paid Search Advertising.
2a. Impact on SEO
Studies indicate that features like featured snippets, which are similar to AI overviews, can reduce click-through rates (CTR) on organic search results. For instance, an Ahrefs study from 2017 found that when featured snippets appear, the CTR for the first organic result drops from 26% to 19.6%, with the snippet itself getting 8.6% of clicks (Ahrefs: Impact of Featured Snippets on Click-Through Rates). This suggests AI tools providing direct answers may similarly decrease clicks on ads and organic links, supporting your observation.
“Evolving SEO for 2025: What Needs to Change” discusses AI’s central role, with Google AI Overviews launched in May 2024 and throttled back due to issues, now resembling traditional SERPs. It notes large language models (LLMs) like Perplexity and ChatGPT synthesize multiple sources, and social search (Reddit, X, Quora, Medium.com, TikTok) integrates into results. The focus shifts to user-centric strategies, predicting LLM platforms will gain traction and community engagement will drive growth, aligning with the need for adaptive SEO.
Optimizing for Google AI Overviews
While the direct Forbes article URL was inaccessible, a similar piece “Google’s AI Overviews: What Marketers Need to Know” explains AI Overviews as AI-generated summaries above search results, potentially eroding organic traffic by reducing clicks. It suggests optimizing for concise, authoritative content, with implications for SEO strategies, supporting the hypothesis of decreased click-through rates
Content Creation Strategies
Content must be high-quality, authoritative, and user-centric to optimize for LLMs and AI engines. The following strategies are recommended:
- Comprehensive and Authoritative Content: Create detailed, in-depth articles that cover topics thoroughly, ensuring accuracy and currency. For example, a fitness client could develop a guide like “How to Build Muscle,” backed by scientific research, using natural language to appeal to users and AI.
Building on this, this concept substantially overlaps with entity SEO, ensuring that content and keywords are well defined to prevent misunderstanding and relevance mismatch. For example, optimizing for the keyword “apple” by providing a broader context ensures that Google, through Natural Language Processing, has a significantly better understanding of your topic. More helpful reading: https://www.clearscope.io/blog/what-is-an-entity-in-SEO
- Adding citations, direct quotes, and recent statistics can boost visibility in LLM results by over 40% in AI responses (arXiv).
- Natural Language Optimization: Craft content conversationally, avoiding jargon unless necessary, to align with how LLMs process human language. For instance, explain fitness concepts in simple terms for broader accessibility, as LLMs excel at synthesizing such content, as seen in Google’s AI Overviews.
Example:
- Topic: “How LLMs Prioritize Content”
- Example: “Think of LLMs like a librarian who categorizes books by subject and relevance rather than title alone.”
- Keyword Strategy: Include relevant keywords naturally, focusing on context and meaning rather than density, to enhance AI understanding. For example, use phrases like “best exercises for weight loss” within a broader discussion, ensuring alignment with user intent, as noted in Optimizing for Google AI Overviews.
- When you write about weight loss exercises, utilise words that relate to that, such as typical exercises, things not to do, as well as the broader topic around weight loss. This will help you both build credibility and authority.
- Encourage Engagement: Foster user interactions through comments, shares, and social media to signal content value, which AI algorithms may prioritize. For example, include calls-to-action like “Share your fitness journey below,” as high engagement metrics like time on page and low bounce rates can influence AI visibility, per Transforming Search: The Impact of AI on User Behavior.
Practically, this may mean you need to review your existing content for “keyword stuffing” – high levels of the same keyword in your content for the purpose of increasing rankings – and rewrite content for natural language (conversational style) combined with increased external direct content to improve credibility and authority.
Secondly, allowing users to engage with your content through comments, polls, uploads is a great way to enrich content with user-generated content (LLMs love this as it’s conversational and not typical SEO-focused content)
Diversifying Content and Brand Building
To further enhance AI visibility, diversify content types and build a strong brand presence:
- Varied Media Types: Include videos, images, and infographics, as AI tools increasingly process multimedia. For example, a fitness video on “Top 5 Weight Loss Exercises” can enhance engagement, aligning with Video-First Content trends.
- Brand Presence: Establish trust through consistent, high-quality content and community engagement on platforms like Reddit and TikTok, which AI considers for responses. For instance, share fitness tips on social media to build authority, as noted in Social Search Brand Engagement.
Website Structure and Technical Optimization (aka Technical SEO)
Technical optimization is crucial for ensuring content is easily accessible to AI tools and search engines. The following practices are recommended:
- Ease of Crawling: Ensure websites are mobile-friendly with fast load times, using clear navigation and internal linking. This helps search engines and AI tools index content efficiently. For example, link related fitness articles to improve discoverability, as noted in Google AI Overviews Optimization Insights.
- Structured Data Markup: Use schema markup to provide context, increasing chances of appearing in AI overviews. For articles, use Article schema; for guides, use HowTo schema.
Example:
<script type=“application/ld+json”>
{
“@context”: “https://schema.org”,
“@type”: “HowTo”,
“name”: “How to Build Muscle”,
“description”: “A step-by-step guide to building muscle effectively.”,
“step”: [
{“@type”: “HowToStep”, “text”: “Set clear fitness goals.”, “position”: 1},
{“@type”: “HowToStep”, “text”: “Develop a workout plan.”, “position”: 2},
{“@type”: “HowToStep”, “text”: “Focus on proper nutrition.”, “position”: 3},
{“@type”: “HowToStep”, “text”: “Ensure adequate rest and recovery.”, “position”: 4}
]
}
</script>
Documentation for schema can be found at Schema.org. This markup helps search engines understand content structure, as highlighted in AI Overviews: Evolution of Google SGE.
- FAQ Pages: Create FAQ pages with structured data to answer common queries, increasing visibility in AI responses. Example
This can enhance visibility in AI-generated responses, as noted in Google AI Overviews: Everything you need to know.
Backlinks
As mentioned earlier, backlinks are a strategic area to evaluate to ensure both your backlink domains and the content that’s published there. Platforms that utilize more conversational-type content combined with user-generated content that gets upvoted and is community-controlled, like forums, etc., are strong signals for LLMs compared to, e.g., Google’s original backlink profiles based on DA scores based on traffic and referring domains.
Examples of strong signals are platforms like Quora, Reddit, X, Medium.com, and similar forums. You can review your website’s backlinks using tools like Semrush and its backlink audit. Below is an example of the bbc.co.uk domain, which could clearly benefit from more UGC content and multimedia content to significantly increase its eligibility into LLMs.
More on how to do all of the above is in section 3 of this article.
2b Impact on Google Advertising
Although Google has been using ML and AI in their Ads for a number of years now, over the past 18 months, there have been notable shifts in the way Google has started incorporating and priming advertisers to trust their AI-based advertising a whole lot more.
Whether it is AI-based auto-applied recommendations, pmax, demand gen, broad max, AI-based shopping optimizations, or Generative AI-based content recommendations. As you can see, the list is getting long with a number of benefits for efficiency, yet, in our experience, mostly at the detriment of quality and control within the context of the specific advertiser in its specific market and budget.
One of the key trends we’ve seen is in the relationship between impressions and click-through rates on search ads. For a number of industries we operate in, we’ve seen this trend:
- Ad impressions up 14%
- Clicks down 3%
- Queries up 9%
- Ad impressions up 32%
- Clicks up 19%
- Queries down 8%
- Ad impressions up 13%
- Clicks – 0% change
All of these scenarios from different industries show an increase in ads being shown while engagement (clicks) with these ads has dropped. In some scenarios, impressions have even gone up while queries (demand in the market) have gone down.
The evidence leans toward a reduction in click-through rates (CTRs) on ads and organic results, with AI Overviews correlating with a 12 percentage point decrease in paid CTRs, as noted in How AI Overviews Are Impacting Paid Performance.
A major reason for this is Google’s push into Pmax and demand gen campaigns and, even more so, its increasing the number of signals it now uses to match potential customers with your ads.
The above example shows Google’s tendency to want to push the boundaries, in this case by including the finance option without having this specifically in the search themes that were inputted. For this business, this is a relevant match; however, it requires manual intervention to prevent matches like this if they are not wished for.
A big one that has stood out recently for me is Google’s liberty with keyword match types. Although I don’t mind using broad match in certain campaigns, Exact and phrase match were always my go-to for curtailing Google’s “spray & pray”.
Where phrase match, for e.g., “brown leather couch,” would always make sure these words would all be in the search term and in that order, that level of control is now also part of the Google Ads graveyard. Now, only the exact match gives you the kind of control that the phrase match used to give. See below table.
In essence, this means that exact matches are only possible by using phrase match combined with a really strong negative keyword strategy.
Google Ads Search Max Match Type
The article “Google Ads Search Max Match Type” discusses a new feature combining search term matching and optimization, aligning with AI-driven trends. While not directly about user behavior, it indicates Google’s adaptation to automation, suggesting advertisers must adjust strategies, indirectly supporting the need for SEM changes due to AI.
This has more recently amalgamated into Google’s launch of AI Max, which basically refrains from manual keyword input and rather builds ads, keywords, audiences, and landing pages based on what you already have live. It is even more important to have your current setup 110% ready.
In summary, Google’s increasingly broad reach to match signals with searches is not going away. It may, and will most likely, improve, but it’s here to stay, and we must adapt, too. Here’s what worked for us in order of priority.
- Implement an advanced measurement setup by utilizing first-party data in your campaigns. For example, we were able to improve lead quality by 46% by integrating with a client’s CRM and optimizing for validated leads instead of all leads, thereby significantly improving closing ratios and sales while not increasing ad spend.
For retail/ecommerce, this means implementing profit-based product advertising instead of just focusing on revenue. - The basics of a good Google Ads account remain strong to generate the best results. Isolated, relevant keywords with matching ad copy and landing pages are still the best way
- In addition, utilize Pmax, demand gen, broad match, and DSA to mine new search terms that work well for you and isolate those into dedicated search campaigns for optimal results and quality scores. This article gives a great example of how to implement this practically.
- We should have plenty of creative assets (images and videos) to experiment with and give the algorithms enough to work with and match.
Summarize this article in ChatGPT
3. Next Steps
a) Measurement
You can use a combination of general website auditing tools and manual checks to review how well your website is doing in terms of being incorporated into Large Language Models (LLMs). There is, however, a great way to now assess the full scope of your website for LLM readiness in a simple way.
Prebo Digital has built a platform that audits your website for LLM readiness and eligibility in a matter of seconds across all the factors we’ve discussed in the above article.
If you want to click through to the audit platform immediately, click here.
If you want to see what it looks like, here are a few examples with sample data.
Using General Website Auditing Tools
Research suggests that tools like SEMrush (SEMrush), Ahrefs (Ahrefs), or SERanking (SERanking) can help. These tools check:
- Technical aspects like crawl errors, page speed, and mobile friendliness.
- Content quality to ensure it’s comprehensive and up-to-date.
- Backlink profiles to assess your website’s authority, which can influence AI visibility.
SEMrush has a great tool that shows you how much your site is already being indexed for particular queries:
The above example is from bbc.co.uk, showing the significant potential awaiting them for further inclusion into AI overviews. Going deeper into this, you can see specific examples of queries, volume, potential links, and current SERP results.
The below example for the keyword “medal table” gives great insight into its current inclusion and a preview.
Continuous evaluation of these types of reports based on search volume, content on your site, and technical checks will greatly enhance your eligibility for LLMs.
More helpful details on this link: https://www.semrush.com/kb/1435-ai-overview
Checking Structured Data
Use Google’s Structured Data Markup Helper (Google’s Structured Data Markup Helper) to ensure your website has proper schema markup, which helps search engines understand your content better and may improve LLM inclusion.
Analyzing AI Referral Traffic
Set up Google Analytics to track traffic from AI tool domains like chat.openai.com or bard.google.com. If your website gets traffic from these sources, it suggests your content is being used by LLMs.
Manual Testing
Perform searches using AI search features like Google’s AI Overview (Google’s AI Overview). Enter queries related to your content and see if your website appears in the AI-generated responses. This can give you a sense of visibility.
By combining these methods, you can get a good picture of how well your website is positioned for LLM incorporation, even though no single tool is designed specifically for this purpose.
Author
Timo Dinkelman, CEO of Prebo Digital, embodies passion and relentless motivation for growth in all facets of life. With a robust entrepreneurial mindset and a focus on digital marketing excellence, Timo partners with brands to drive business growth through proven strategies and a strong partnership approach.
His business acumen is backed by qualifications from VU University Amsterdam and a five-year tenure at Google, where he achieved top-performer status in EMEA. Since founding Prebo Digital in 2017, Timo, together with his partner Precious Thundu, has led the agency to significant success; Being a Google Premier Partner since 2022, and one of the first verified Amazon Ads Partners in South Africa, supporting multiple businesses to reach their marketing objectives.
4. Key Citations
- ChatGPT Growing as Traffic Referrer, Changing Search Behavior
- [AI Search Gaining Traction, Not Replacing Google: Survey](https://searchengineland.com/ai-search-gaining-traction-not-replacing google-survey-451667)
- Evolving SEO for 2025: What Needs to Change
- Google’s AI Overviews: What Marketers Need to Know
- Google Ads Search Max Match Type
- Google’s AI Search Feature Fuels Content Traffic Concerns
- Search Powered by Chatbots will Impact Your Search Strategy, You Ready?
- Impact of Featured Snippets on Click-Through Rates Study
- Featured Snippets Study: Results From 3,500+ Internet Users
- https://www.semrush.com/kb/1435-ai-overview
