While Google SGE has still not been deployed in Europe, 2024 should see it arrive on the French internet. However, its arrival will not only have consequences on the search habits of Internet users , but also on the natural referencing (SEO) of websites.
So, to keep up with you and prepare for the arrival of SGE now, let's recap everything you need to know about this new tool , from its features to its limitations, including what it means for SEO professionals.
In this photo, the facade of the Google company.
What is Search Generative Experience, or SGE?
First thing to know: Search Generative Experience (SGE), at least as described by Google last May, is shaping up to be a significant step forward for the Web. This innovative concept, revealed at the Google I/O conference dedicated to developers, is indeed likely to usher in a new era for the user experience on the Internet.
To put it simply, SGE is a conversational interface based on generative artificial intelligence . It aims to radically transform the way Internet users interact with search engines, especially since Google wants to make this functionality very accessible, both on mobile and desktop.
Concretely, when the user formulates a query in natural language, the SGE will offer him a sourced, precise and personalized response . This will be distinguished at the top of the results page, in a colored insert placed between the sponsored results (SEA) and the natural results (SEO).
In practice, Google SGE will allow information to be gathered into a concise summary , which will promote a faster and more comprehensive understanding of the subject being searched for. Through this strategy, the American giant would therefore evolve towards a role of "answer engine" , to the point of potentially reducing the need for users to visit other websites to obtain additional data.
It goes without saying that SGE is a major step in the evolution of content marketing and SEO, as it will deliver a profoundly improved and more intuitive search experience .
Webinar
How to reach the first position on Google?
Watch the webinar!
How Google's Search Generative Experience Works
Google's Search Generative Experience (SGE) is expected to skype database offer a series of features designed to improve the user experience on the web. Additional questions, image generation, contextualized translation, etc., we take stock of all the promises of the SGE.
Ask additional questions
One of the key features of the SGE is undoubtedly the ability for users to ask additional questions. This function, called Ask a follow up , should allow you to refine any search by means of an additional query. Its effectiveness is notably due to the fact that the tool adapts in real time to the Internet user and their intention , for example by retaining the context of previous questions. This also applies to the links communicated under the answers, since they will be updated as the conversation progresses.
Generate images
It goes without saying that Google SGE should integrate an image-generating artificial intelligence , similar to Midjourney or Dall-E . It will therefore be possible to produce visuals in a few clicks, thanks to a query entered in the Google search bar or directly in Google Images.
Promoting contextualized translation
To further break the language barrier, Google SGE should include a contextualized translation feature, based on the same principle as applications like Google Translate or DeepL . The tool will be ideal, among other things, for displaying significant differences in meaning between words and allowing the Internet user to choose the most appropriate meaning.
Show definitions
For technical terms , Google SGE will allow users to access a preview of certain definitions. Visuals or images related to the query may also be offered to facilitate understanding by Internet users.
Generate a product list
As soon as a query includes a purchase intention, the SGE will be able to generate a list of products with reviews, ratings, prices or even photos. In addition, depending on the context, the latter should be able to be geolocated and adapt to the user's local constraints.
Show sources used
To facilitate the traceability of the information provided, Google SGE will clearly indicate the sources used in the summary provided for this purpose. It must be said that contrary to certain preconceived ideas, these do not always correspond to the first organic results. The user will thus have the assurance that the information provided is reliable and that it comes from serious resources.
What are the limits of the SGE?
As you can see, SGE technology is about to revolutionize the way users interact with search engines. However, like any innovation, it has its own limitations, especially when it comes to artificial intelligence and large-scale language models (LLM). But what are they? That's what we're going to see.
Misinterpretations
Despite the rapid evolution of technology, artificial intelligence does not always respond correctly to the queries made to it. This is also one of the criticisms that frequently comes up about ChatGPT and Bard . The challenge for Google SGE will therefore be to limit the risk of misinterpretations when corroborating data. Indeed, although the tool is capable of correctly identifying information to provide an answer, subtle errors (interpretation of language, etc.) remain frequent and are likely to alter the quality of search results.
Inventions and erroneous facts
Inventions , erroneous facts or mixed data form another limitation observed with LLMs, including in the case of Google SGE. Bard is for example the champion of invention of all kinds, and even more so in the case of geolocated data. When the model does not have clear answers to provide, or when the query is too complex, it generates misleading content . Thus, if you are looking for information on creative workshops for children in the Oise department, it is a safe bet that you will be directed to workshops in other departments, and this, without this being explicitly told to you.
Also note that biases may remain due to the data on which the SGE model was trained. If the data reflected limited representations or implied negative contextual associations, the tool is likely to produce biased results .
Contradictions that can confuse the user
The integration of SGE into Google’s results pages can potentially create contradictions. Indeed, where a featured snippet presents a single perspective, SGE offers a synthesized answer that draws on a range of sources. This situation can obviously generate some confusion among users , to the point of prompting them to launch into more in-depth research on their own.
