What is an AI chatbot?
To better understand how to optimize a website to perform well for AI search results you have to know what an AI chatbot is. An AI chatbot is essentially like any other chatbot you have used in the last few years but instead of utilizing a list of canned responses based on certain criteria it utilized artificial intelligence to provide the answer. This move from canned responses is where the evolution to generative AI comes into play.
Basically it is a system that mimics human conversation utilizing more conversational AI techniques and natural language processing. This allows for much more sophisticated results with greater detail and, hopefully, better accuracy while giving the user a sense they are actually speaking to a human. This conversational aspect is what creates such an interesting experience as the chatbot adapts in its tone and refines how it provides information
The rise of AI SERP features
AI Finds Its Way into SERPs
As soon as ChatGPT came onto the market in November 2022 the digital marketing world knew it wouldn’t take long for generative AI to infiltrate our spaces, especially SEO. Countless companies, including Google and Microsoft, began experimenting with their own generative AI experiences that sat outside of traditional SERPs.
As time moved on, Bing began incorporating gen AI into their results utilizing their Copilot product. Google also began bringing AI into their search experience with the inclusion of the AI Overviews SERP feature which is powered by their Gemini LLM in March 2024.
The AI Overviews feature, in my opinion, was the first time SEO really felt an immediate shakeup from AI. At time of writing this, it is being reported by Ahrefs that the AIO feature is showing 116% more often than it did upon original release. This has caused an estimated -34.5% drop in click-through rates on organic results in position one.
The consensus across the industry is that since the inclusion of AI Overviews into the SERPs clicks have gone down causing a drop in traffic while the number of conversions from incoming traffic is going up. This tells us that a pivot in what we consider success metrics from SEO needs to evolve. The traffic coming to sites is likely more qualified so SEO needs to work closer with UX and design teams to make sure the experiences we are pointing our traffic to feel as little friction as possible to help improve conversions.
Common Query Types and Citations
At the time of writing this, Semrush is reporting that 13.14% of all search queries trigger an AI Overview feature in March 2025 which is up 6.49% since January 2025. 88% of those queries lean more towards informational, top-of-funnel results. So, as you can imagine publishers who focus heavily on this content type are less than enthusiastic about this impact on traffic to their sites.
While the top-of-funnel content is the primary concern at the moment, more commercial intent queries are starting to see AI Overviews. AI chat platforms themselves are beginning to utilize AI Agents to help users make purchases and perform buying actions. This sort of action is going to depend greatly on user trust so we will see if it becomes more common for users in that buying stage and how that will impact SERPs.
While we don’t have the ability as SEOs to fight this new reality we can certainly make the best of it. By optimizing our content to be friendly for LLMs to understand we can gain more active citations in generative AI results and hopefully continue to see the traffic benefits we aim for while continuing to build up brand awareness within AI results
How to track performance
Unfortunately, because these platforms are so new, analytics platforms like Google Analytics 4 (GA4) do not track AI channels natively. That being said, it’s fairly straightforward to set up views to help get a better understanding of what traffic is coming to a site through an AI system.
Custom Channel Groups
One of the first things I like to do is establish a custom channel group so that I can use it to segment the data in my reports. Doing this allows you to plot charts and pull comparative data directly in GA4 to compare channels against each other, like organic search traffic, paid media, social, and AI platforms.
GA4 Explorations
Another great way to get more data on AI traffic is to establish an exploration view in GA4. My favorite aspect about this solution is that you can see the exact referrers driving traffic from AI platforms. This allows you to see what type of content is being favored by LLMs and then making decisions on how you write and present your content to get even more visibility in these result types.
Setting Up Reporting
I certainly don’t want to take credit for other people’s hard work but one of my favorite instructional pieces of content on setting up custom channel groups and explorations comes from the team over at Obrit Media. I highly encourage you to check out their guide for getting this set up in your GA4 account to have even better reporting from AI traffic.
Google Search Console and Bing Webmaster Tools
While we can manipulate analytics platforms to see traffic from AI platforms we are missing this data from tools like GSC and BWT. It is contained within the existing data but not in a way that we can segment and get insights to make decisions again. At least not yet. SEOs continue to beg Google and Bing employees to provide this data so hopefully one day we will see it in their platforms.
Prospecting new optimization opportunities
To make the most out of AI search and SERP features, it is important to be able to utilize the tools you already have to find new opportunities to get ahead of your competitors. One of the best ways to do this is by using your keyword and site research tools to see what you are performing well for in SERPs, including SERP features like AI Overviews, and then finding where you could be doing better to build an optimization strategy.
Here are a few common tools in most SEO toolbelts that can give you an idea of where you can make improvements to increase visibility in AI search.
Ahrefs
Within the Ahrefs site explore tool is a report that I love to use to determine what opportunities I have to gain more momentum in AI results. This will give you an overview of the entire site, subdirectory, or page you are reviewing while showing what terms are showing an AIO result and if you are ranking or not.
Steps to reviewing this report
- Input your site URL into the Site Explorer tool in Ahrefs
- Navigate in the sidebar menu to the Organic Keywords report
- In the filtering options at the top of the report, select SERP Features
- Then, select the AI Overviews feature and check Where target doesn’t rank
Once you have your report filtered you will have a better understanding of what keyword opportunities you can go after to improve existing content to get more visibility within AI Overviews. I usually like to go after the keywords and pages that are not currently on page one as these improvements have a dual benefit through the work you put into them.
SEMRush
Similar to the report I previously mentioned from Ahrefs, Semrush can also provide the same type of data. You can put a website into their Domain Overview tool, review the positions the URL ranks for in SERPs, and then filter the table to only show results with AI Overviews.
Steps to reviewing this report
- Input your site URL into the Site Explorer tool in Ahrefs
- Navigate in the sidebar menu to the Organic Keywords report
- In the filtering options at the top of the report, select SERP Features
- Then, select the AI Overviews feature and check Where target doesn’t rank
Utilizing this report will give you the insights you need to prioritize your optimization efforts to improve visibility within AI Overviews. Plus, just like the previously mentioned report, you can prioritize based on what keywords are not on page 1 to get the most impact from your efforts.
seoClarity
Another great tool to utilize to track your performance in AI overviews is seoClarity. Within their Research Grid tool, you can filter to show just the keywords showing within AIO. This gives you an idea of what is already performing well while giving you an idea of what might be lacking so that you can make decisions on prioritizing optimization initiatives.
In addition to the Research Grid tool, within the Rank Intelligence tool you can see trend data on daily performance within AI Overviews as well as what other competitors were cited in those results. This gives you more competitive information to help influence your optimization decisions to maintain your advantage over competitors in your site’s vertical.
What kind of optimizations help AI
From what most research is showing at the time of writing this post, most standard SEO initiatives can make a big impact on a site’s appearance in AI search. That being said, there are a few things you want to make sure you focus on to give your site a leg up.
The main consideration when making optimization decisions for AI search is to pay attention to the entities across the topics you are producing content around. Entity optimization seems to be the key to the puzzle of performing well in AI search. So make sure you are doing your research to build out logical topic silos while utilizing technical SEO best practices to make sure those entities connect to each other correctly.
Then, as you are creating content you are writing it in a way where an LLM can process the content and understand how the words you have written relate to each other. Essentially, you want to make sure that the vectors established in your content (the entities) establish a logical embedding (the semantic relationship between words). This helps the LLM act similarly to a human brain and return results appropriately when related to your site and the topics you create content around.
Structured Data Markup
There has been some discourse over the last few years on the importance of structured data to SEO now that search engine crawlers have become so much smarter and efficient. While I can see the argument for not spending time adding something that makes an immediate and measurable impact to a website, I think that mindset is a bit dangerous.
I personally see SEO, especially technical SEO, as throwing everything you can at the crawlers to give yourself an advantage. Structured data markup is one of those activities that might be difficult to measure the success of immediately and attribute to revenue but it can help the entire ship float more efficiently so it is worth the effort, in my opinion.
Luckily for all skeptical SEOs, we got a better understanding of the importance of structured data from Fabrice Canel, Principal Program Manager leading the Bing Web Data platform team, on how Bing/Copilot utilizes structured data to influence its results. Essentially, according to Canel, structured data markup helps LLMs, specifically within Microsoft, better understand the content of a page and its overall freshness.
I would say, from my personal experience, utilizing some of those key markup types outlined by the Google Developer Guidelines would be helpful to not only traditional search and acquiring SERP features is a solid investment. Adding this information to your site will give LLMs and crawlers a better understanding of the information you are providing them.
Internal Linking
If you think of a website as a house with each page being represented by a room, internal links are the hallways that allow you to get from room to room. Without these hallways, users cannot get to the rooms they are trying to go to as their end destination or even explore the house further.
The same can be said for an LLM. To better understand a website and the relationship between pages internal links are crucial. That navigation allows a crawler or users to explore a website further and consume more relevant content. From a technical perspective, this helps with the flow of equity from page to page which can have a big impact on rankings.
Technical SEO
Technical SEO acts as the foundation for the metaphorical house a website represents. It keeps the house standing sturdy so that the inhabitants of the house can move through the home to meet their goals. At the end of the day, excellent content can only get you so far as it is only one piece of the puzzle.
When considering technical SEO outside of the elements we’ve already mentioned, always consider how a crawler and LLM interacts with the site once it gets there. Ask yourself the following:
- Does this website load quickly?
- Is it rendered completely upon initial load?
- Are there any elements blocking content from being readable?
- Is it easy to navigate between pages?
If you can confidently answer those questions your website should be in a position to perform well in SERPs and AI results/features.
The future of SEO and AI
While it is easy for folks to say that “SEO is dead” it is more important than ever. The work is just evolving like it always has. If people are still searching there will be a need for SEO.
At the end of the day, with the current iteration of AI and how these platforms return results we are realistically in a “business as usual” mentality with our work. Things are evolving and can change quickly but as long as we are making smart decisions with our SEO initiatives we should see great results. Just don’t be afraid to test new techniques and strategies to see what really works for the vertical you focus on and pivot accordingly. Happy optimizing!
