In the evolving world of SEO, staying ahead means adapting to new technologies like Google’s AI Overviews and OpenAI’s ChatGPT.
This article explores how these tools function, offers strategies for optimizing your site today and anticipates future changes in search engine technology and user behavior.
For many SEOs, Google remains the loudest voice in the room. Automation has long been a part of what makes a Google search effective, but the face of it has recently changed with AI Overviews.
Google’s AI Overviews show summarized information directly on the search results page. They aim to answer questions without needing to click on other websites. It’s like a “live” featured snippet that gives quick answers and includes links for more details if needed.
Notably, the percentage of SERPs showing AI Overviews has fluctuated as the public reacted to its release. While users found many “errors,” Google pulled back on the frequency of its appearance, though not to zero.
Beyond Google, there’s a competitor trying to steal the scene: ChatGPT.
Though there are tools offering similar services, the undeniable popularity of ChatGPT as a household name created urgency and big questions:
When Google performs a search using its algorithm and determines it can offer the user a summary, the search engine accesses a site in much the same ways it would to create a SERP appearance.
From a technical perspective, optimizing for Google AI Overviews is a process that largely looks like optimizing for Google Search.
When it comes to keyword strategy, site owners should continue tracking queries for popular topics, paying attention to how the brand is represented in terms of market share, brand voice and informational accuracy. As always, answer questions with helpful content. E-A-A-T is paramount.
Until and unless Google releases a way to track the appearance of AI Overviews on a SERP, SEOs will have to rely on manual testing or makeshift approaches to understand how often AI Overviews appear and for what types of queries.
Dig deeper: How will generative AI impact website rankings and traffic?
The difference between a traditional search engine experience and one built around a large language model (LLM) in a chat-based platform is that the search moves away from “browsing a series of links” to a starkly clean conversational space.
Today, ChatGPT comes in a few flavors, most of which can access the web, especially if prompted by the user with inputs like “look up,” “search online for” or “cite your sources.”
When the tool doesn’t connect to the web, it instead predicts a likely response based on its training materials.
Dig deeper: What is generative engine optimization (GEO)?
There are two scenarios:
Using Microsoft Bing, ChatGPT will “search the web” like a human user, reviewing search results pages and even visiting websites for a quick read when enabled.
Links are provided as citations and sometimes recommendations. Users can click to see the code behind the search, too. However, it is unlikely that the average person would care to do so.
In other words, if people like quick answers without clicking through several links, they may prefer a tool like ChatGPT.
Optimizing for this would be like optimizing for any search engine, though it might be required to allow the tool to access your site.
Due to concerns about losing traffic or people accessing content within the tool, some site owners are currently hesitant to open the door to ChatGPT. This protects content but limits visibility, giving it away to counterparts not afraid of the risk.
Dig deeper: 3 reasons not to block GPTBot from crawling your site
Many inputs in LLM chat tools, especially if not specifically stated, will draw a general text prediction response without looking to online sources, instead drawing on “predictive text.”
In this case, one cannot optimize for clickable appearances at all.
Information about a brand or business must exist within the bots’ training texts. Debates continue about which source texts exist, when they were accessed and whether they were ethically or legally sourced. OpenAI maintains documentation on how ChatGPT is trained.
Partnerships with sites containing a large body of conversational data (e.g., Reddit, Quora, etc.) may offer LLMs information without links. Optimizing for them can become a question of public relations – where are people talking about the brand or business, and what are they saying?
In effect, building positive interest and limiting “bad press” on the web will come down to actions taken in real life, which will naturally reflect in digital spaces, not the reverse.
Organic clicks and impressions, including engagement, visibility, CTR and average ranking, have traditionally been key SEO metrics. But things are changing.
It’s impossible to know whether a click in Google Search Console originates from a traditional “blue link” click by a human user or from AI Overviews. This lack of clarity extends to the quality of traffic generated by AI tools like ChatGPT.
For instance, when an AI-generated summary answers a query without resulting in a human website visit, creating a “zero-click” situation, traditional success metrics won’t apply.
Also, if the search happens on a platform other than Google, it won’t show up in Google Search Console, making it hard to measure SEO success in an AI-driven world.
Note: Referral traffic from ChatGPT can be tracked in Google Analytics, providing a partial solution.
Dig deeper: Could AI eventually make SEO obsolete?
