The AI revolution is well underway and transforming the digital landscape faster than you can say “Screaming Frog.” Now, SEOs are facing a lot of unknowns, particularly when it comes to AI’s impact on SEO. It seems like just yesterday we were grappling with the changes brought on by Google Analytics 4. Now, it feels like our whole world is on the verge of flipping upside down.
I recently took a deep dive into Large Language Models (LLMs) such as ChatGPT, Bing Chat, Claude, Bard, and a bunch of plugins and extensions. Now I’ve been thinking about the impact these tools are going to have on our SEO performance metrics.
What does this mean for performance accountability? How can we optimize sites for AI results? How will we accurately measure and report performance? What impact will there be on keyword usage?
This article is my attempt to shed some light on these questions. For a more comprehensive understanding, explore my guide on ‘AI in SEO: Essential Guide and ChatGPT, LLM Strategies’.
AI’s Impact on SEO: Will Click-Through Rates Drop?
Search engines will be integrating AI directly into their search results. Google will have an “AI-powered snapshot” at the top of the SERPs page any day now. We now know is called the Search Generative Experience (SGE). Bing will have something similar. I wonder, what does this mean for how people find the information they are looking for? If they find their answer in the AI results, they may not feel the need to click through to a website, potentially lowering click-through rates.
For businesses that rely on traffic to provide revenue, how can they address this challenge? I believe businesses will have to emphasize high-quality, in-depth content that goes beyond quick answers.
However, adapting to this new landscape involves more than just improving content quality. It requires a strategic understanding of how AI like Google’s SGE potentially affects search behaviors and SEO practices. To learn more about effectively preparing for and navigating the challenges posed by Google’s SGE, dive into my article on preparing for Google’s AI-driven Search Generative Experience.
Search Engines as Information Providers
Relatedly, as AI continues to evolve, people will more commonly find the information they need directly in Google, Bing, and other search engines. I think this will change the search engine’s role from being merely a conduit to the desired information to becoming the primary information provider. Businesses will have to alter their strategies to focus more on visibility in answer boxes or snippets.
I think the focus for businesses and other content creators will be on taking up as much real estate on search result pages as possible. You will have to do this by gaining visibility in LLMs and AI. The only way to do that is by blending your SEO and AI optimizations.
Content Selection for AI Output
As of today, there is very little clarity on how LLMs are selecting information for their outputs. LLMs are built on digital training data encompassing publicly available information on the internet. For example, ChatGPT 3.5 was trained on internet data up until December 2021. Both Bing Chat and Google Bard have access to the current internet to pull information from.
The criteria for how the LLMs evaluate material for outputs are still in the dark. I wonder if LLMs will favor older sites for information as they are likely rated as more authoritative on the subject. For newer businesses, I think the emphasis should be on originality, in addition to high-quality, up-to-date content.
Accounting for AI in Search Analytics
Understanding how Google Search Console or Google Analytics will account for AI-generated search results is a must. At present, the methods search engines will use to categorize and track AI-generated content remain uncertain.
When Google search includes your content in an AI result, will they count it as an impression? Will they include that in the existing impressions or separate it into a new category of AI impressions? Will they count a click in an AI result as an existing click or classify it as an AI click? How will they account for this in the click-through rate?
Although I expect analytics tools to evolve and provide more granular data, there is no clarity as to how or when this will happen.
Optimizing Websites for AI
In the process of optimizing websites for AI, developers trained LLMs using a diverse range of data, which included information from both current and older websites. This information can be out-of-date or just plain wrong.
However, I believe this approach introduces a potential bias as the LLMs might lean towards favoring established websites that possess a substantial volume of content. Consequently, these websites are likely rated as more authoritative and trustworthy by the AI systems.
Focusing on Newer, Unique, and Trending Ideas
I think this could place a premium on the creation of unique, current, and trending content that may not be a part of the AI’s training data. Businesses will have to stay ahead of the curve by differentiating from AI-generated content. They will have to focus on producing more unique content that taps into current trends and ideas.
The Future of Keyword Optimization for AI
How we optimize keywords may also see changes in the age of AI. As AIs become more adept at understanding language in more nuanced ways, the focus may shift from specific keywords to more contextual, topic-based content. Or, I think we may see a change to more conversational phrases rather than keywords as AI becomes more focused on speech rather than textual input.
Google’s Stance on AI-Generated Content
How will Google and other search engines handle AI-generated content? With a flood of AI-generated content coming, search engines will likely have to find a way to limit the scope in one way or another. Will search engines favor AI-generated content that is more accurate and concise? Or will search engines disfavor it due to potential issues with originality? These questions SEOs will have to closely monitor. In the meantime, I suggest that SEOs should continue to create AI content responsibly to stay ahead of the curve.
Search Filters for AI Content
As AI-generated content proliferates, I wonder if search engines may offer filters that allow users to choose whether they see AI-generated content. This can have a sizable effect on SEO. We may have to develop separate strategies for AI content and human-generated content.
Conclusion
As we continue to try to stay ahead of and predict AI’s impact on SEO, I believe SEOs must stay informed and flexible. The future of AI may be unknown, but it’s undoubtedly exciting. Read my latest post if you might be wondering, “Is SEO Dead?”
