by Kerry van der Mescht
Google is in the testing phase of a new way to generate and display search engine results. Called Search Generative Experience (SGE), the new approach uses generative AI to deliver more meaningful search results to users' queries.
Google's AI model allows users to have more natural conversations with search engines while also offering creative features to help users apply the information they just learned.
Google has been using a "lite" version of AI for years to deliver relevant search results to its users; now, it is using generative AI that can create new content—search results—based on the users' inputted data, search history, and online sources and references.
SGE uses generative AI to change the way search results are delivered, by providing the user with more information from a range of sites in the search results. The goal is to enable users to ask more conversational-type questions and get a thorough reply on the search results page—without having to click through to a website that may or may not contain the answers they seek.
New devices and Google's advancements are changing the way people can access information, as well as asking questions through not only text but also voice-search.
Google's AI has been trained to corroborate responses from trusted sources and designed to reflect a neutral tone in the results provided. With SGE, people can have natural conversations with Google, discovering more relevant results for their queries, with links to related sites for further navigation.
SGE also uses AI to be creative. The AI can create entirely new data for you (for example, the perfect social media post to highlight your new product or create a new kitchen design based on your search history and preferences). Such creativity in generating new content from scratch allows users to go from just finding information to applying it in their lives or in their business.
With SGE, the search results show a system-generated snippet in the main answer box, with three featured trusted sources to click on that support the generated answer. Google shows those high-quality sites with an image, title, and URL in the answer box.
In addition, there is a toggle expander that enables you to dive deeper into related answers, and these additional responses are also generated by AI. Alternatively, you can scroll down to see the traditional results.
Underneath the generated answer and trusted sources, there are follow-up options that the user can take: specifically, an "ask a follow-up question," which helps users find what they are looking for.
Results will also include customer reviews, ratings, prices, and product images, where applicable and available.
The new search results also deliver answers to follow-up questions, at the bottom, as well as suggest next steps. When you click on the Follow Up button, it allows you to ask even more questions about the topic you are interested in. Google will remember your original question and bring that context into the answer to your follow-up questions.
SGE can use AI to generate new content for you based on your preferences and online tone. This new, creative feature of AI, however, can lead to challenges in trusting your search results and finding a balance between creativity and hard facts.
Another limitation of AI in search is the misinterpretation of facts—that is, certain data can be misinterpreted or identified incorrectly, and so affect the results delivered to the user.
There is, furthermore, the risk of bias. Sites contain their own bias in their content, including high-quality sites that have a high probability of being used in the AI-generated results with Google, thus injecting bias in Google's results.
In the case of YMYL (your money or your life) topics, when the search results could have real-world impact on users, Google will add a disclaimer to those results to emphasize that people should not rely on this information solely for medical advice (for example).
"Just as our ranking systems are designed not to unexpectedly shock or offend people with potentially harmful, hateful, or explicit content, SGE is designed not to show such content in its responses," Google explains.
When there is a lack of reliable, trusted information to respond to searches, Google might not give generative answers. Also in the case of searches that are explicit or dangerous, Google might not give generative answers.
There is a debate on how Google's Bard (now Gemini) chat AI service is going to affect Google search—in that a user can search for information on the chat AI service and get immediate results delivered, with content that answers the user's questions.
However, users of search platforms are used to typing their questions into search bars and getting answers from there, with a range of sites to choose they can choose to click on in the search results.
Along with the issue of habit is the issue of trust—questions whether the answers that chat AI systems respond with are necessarily the best answers.
Still, it's likely that Bard/Gemini and Microsoft's Bing Chat will have at least some impact on search.
Google's AI chat service differs from others in that it can take information from any site indexed on Google and analyze relevant data to provide real-time answers to questions.
From Eyes on News: "Google Bard is a conversational AI chatbot. It operates with machine learning (ML), natural language processing (NLP), and generative AI to understand user prompts. Google Bard also provides text responses."
From Near Media, Google's Sissie Hsiao, VP and GM for Google Assistant and Bard: "Bard will be able to tap into all kinds of services from across the Web," including all relevant content on the platform, as well as Google Images, Maps, knowledge graphs, and even Google Translate.
Bing has incorporated Bing Chat into its search engines—available next to the traditional search bar—as an alternative method to search for information.
From Microsoft: "Bing Chat is built into the browser's sidebar for fast and easy access. If you're using Microsoft Edge, select the Bing Chat icon to open the Bing Chat pane. You'll have access to Bing Chat, Bing Compose, and Insights: three powerful and distinctive AI tools."
Bing Chat is also designed to hold a conversation with a user: When someone asks it a question, the chat system provides an answer as well as suggested follow-up questions.
Also, users can search Bing Chat for information about an image they've submitted and can also create their own images using the Bing Image Creator.
In the current stage, as AI systems are still learning in order to become better able to answer user queries, traditional search engines will most likely remain the preferred search option for most users.
However, once AI systems become more trusted, and capable of delivering highly relevant results, chat AI systems could become the preferred platform to search for information.
Currently, a hybrid approach to searches seems on the rise; people are using search engines for most of their searches while asking chat AI systems for less specific or more subjective information.
Google is still using Domain Authority score, as well as the quality of backlinks, to determine which sites are high-quality online sources. The new AI search results also pull out reviews, so a focus on customer reviews will also help boost the results, as well as help Google see a brand as high-quality.
As with Featured Snippets, the content on a webpage likely needs to be laid out in a way that easily answers user questions so that AI can identify that content and use it in its answer. That content should also focus on longer-tail keywords and provide answers to searches that are typed in or asked—for those conversational-type questions, as well as creating organic answers for chat AI systems to use.
On X, Glenn Murray, @SilverPistol, noted: "MY TAKE: It's like featured snippets on steroids. Source sites appear prominently in the answer. Google wants users to 'dig deeper into those sources'. Our challenge is to be one of those sources. i.e. We have to keep producing quality, helpful content."
In addition, content should be optimized to better answer users' queries and follow-up queries. That means more relevant content should be added to webpage content, with research on common or related queries supporting new content additions to pages.
The more relevant content on the site that answers any questions a user might have would be beneficial to the site performance due to the new "follow up" questions that users are prompted to ask. That additional content should be structured in clear H2 and H3 tags so that the search and AI bots can deliver exact results, generating a more successful searching experience to the user.
All such content additions should keep the goal of SGE and AI in mind—that is, helping users find more relevant results by allowing them to ask follow-up questions, leading them to the most relevant answers to their questions.
Sources:
No, ChatGPT Won't Make Marketing and Search Irrelevant (At Least for Now)
How AI Will Affect Delivering Relevant Search Results
ChatGPT vs. Bard: The Future of Google's Search Dominance
Enter your email address to continue reading
How AI Is the End of SEO... Or Not
Don't worry...it's free!
Sign in with your preferred account, below.
Copy Link
AI
ABOUT THE AUTHOR
Kerry van der Mescht is a search manager at Equator, a full-service digital transformation agency. She has been working in the industry for a decade and specializes in SEO and PPC.
LinkedIn profile: Kerry van der Mescht
You may like these other MarketingProfs articles related to SEO:
Keep me signed in
Sign in with your preferred account, below.
Over 600,000 marketers rely on MarketingProfs for B2B know-how every day. Don't miss out on the latest marketing tips and techniques, delivered right to your inbox.
Subscribe today ... it's free!
Sign in with your preferred account, below.
source
