Google's New Generative AI Search Experience: What Does It Mean for Brands?
On May 10, Google announced its new generative AI search update: Search Generative Experience (SGE). However, the update won’t be fully rolled out to the public until 2024.
Google SGE will use machine learning algorithms to understand users’ search queries, browsing history, and interests in order to deliver more personalized and unique results.
Here’s what we believe this will mean for brands:
A heavier emphasis will be placed on surfacing content that’s more conversational and optimized for voice search.
Content production efforts will need to center around personalization.
Videos, specifically, should be a priority for future content creation efforts.
Thought leadership content will help fill in informational gaps left by AI.
Changes in search behavior will require brands to redefine the KPIs they use to determine “success.”
For brands that engage in local search, it will be crucial to keep a close eye on the performance of your business profiles, including Google Business Profiles (GBPs).
Introducing: Google Search Generative Experience
After months of rumors, Google announced their new generative AI search experience under the codename Project Magi. While this rollout is currently only available to users in Google Labs, generative AI search has huge implications for the way advertising, content curation, and search work for individuals and businesses. In this article, we’ll explore what Google’s generative AI search experience is, how it works, and what it means for the future of search.
Here are the specific topics we’ll cover in this article:
“Generative AI search” refers to the use of generative artificial intelligence algorithms and techniques to explore and generate new content, ideas, or solutions based on given parameters or constraints. Unlike traditional search methods that simply find existing information or patterns, generative AI search takes things to a whole new level by generating brand-new outputs that are totally unique to the particular searcher and their search habits.
Meet Google SGE: The Generative AI Search Experience
While Google currently holds over 90% of the search market, Microsoft’s Bing has seen a 25% increase in monthly searches since its integration with OpenAI’s ChatGPT and ChatGPT 4. In an effort to compete, Google is working on its own plans for developing an AI search: the Google Search Generative Experience (Google SGE).
Google SGE is a generative AI search engine that is expected to take user personalization and AI-generated responses to a search bar near you. Using machine learning algorithms to understand your search queries, browsing history, and interests, Google hopes to deliver more relevant search results. More notably, Google will bring chatbot-like functionality to search to create AI-generated answers to your search queries.
Ultimately, Google’s goal is to make search feel more like a conversation. Much like we already use voice search tools like SIRI, Google Assistant, and Alexa to answer conversational questions, Google’s generative AI search engine aims to deliver succinct AI-generated answers without forcing users to dig through search results.
Using a combination of large language models (LLM) like MUM, rumored to be 1,000x more powerful than Google’s 2019 algorithm, BERT, and Pathways Language Model 2 (PaLM2), Google will dramatically upgrade existing machine learning and natural language processing (NLP) capabilities in one of the biggest updates to date. These upgrades will help search algorithms better understand user queries, deliver personalized results, and give users AI-generated answers to their searches.
AI-powered search is the next logical step: relevant answers that are personalized and adapted to us. It feels like a timely evolution beyond featured snippets or knowledge panels by giving us a personalized answer immediately.
How Does Google’s Generative AI Search Engine Work?
While Google SGE is still in testing, the features Google has shared will most certainly create changes in the way users currently use search. Here are some of them:
More Relevant Search Results
With Google SGE, users get more relevant search results and personalized recommendations that are tailored to their interests, preferences, and search history. This means they spend less time sifting through irrelevant results and more time finding what they’re looking for. Contextual targeting – the strategy of displaying ads based on a website’s content – will become more important than ever before.
A Smoother User Experience
Overall, generative AI search results can help deliver a better user experience by compiling an answer faster than having to read through individual search results. This means users are more likely to find what they’re looking for and have a better search experience. A button below the AI-generated answer allows users to ask follow-up questions for further elaboration and details.
Transactions Straight From Google Search
One of the functions included in Google SGE includes sales transactions straight from the search results page through Google itself. Shoes, airline tickets, and other items relevant to your search may be populated beside AI-generated search results based on your personalized search profile.
Google has already implemented similar features for services like restaurant online ordering, where transactions occur within the Google experience rather than having to send a searcher to a third-party website. Ultimately, this enhanced experience could genuinely help smaller websites compete for e-commerce sales on a leveled playing field, but it may also rob larger sites of regular traffic to offer this type of user experience.
Dataset Powered by Google Search
With the sheer size and popularity of Google as a search engine, the number of queries able to be studied will be unprecedented. Google SGE’s ability to analyze user queries and website content will allow it to train on the largest AI dataset ever seen before. This could allow the tool’s ability to improve exponentially faster than competitors like OpenAI and Bing Search.
Application of Google’s Safety & Quality Protocols
Google is optimistic that it can use its existing content quality scoring metrics to create scalable safety and quality protocols across indexed websites. Much like Your Money or Your Life (YMYL) updates, Google says they will opt out of generative AI search results for topics that could cause harm or simply don’t have enough qualitative data. For now, Google Labs is showing responses for “safer” topics and asking for its users to flag any harmful content to help iron out these issues.
Factual vs. Fluid Answers
Behavioral tests show users are much more likely to trust AI-generated answers presented in a more conversational, fluid style, rather than the rigid, factual styles that AI is historically known for. Now, AI can generate answers in fluid styles with ease, but the bias remains. Google SGE is designed to give answers in a factual style, hoping to make AI content more identifiable to users to signal the necessity of fact checking.
How Will Google’s Generative AI Search Engine Affect SEO?
This will be one of the larger algorithm updates since BERT and RankBrain, but until its widespread launch, we can’t be certain exactly how SERPs will fare. This update may completely redefine what we know about click-through rates using traditional search.
Currently, a traditional organic search result in position #1 accounts for roughly 40% of a query’s clickthrough rate, but that may sharply decrease if a user is able to get their answer without clicking through to the answer source.
Here are some things we can expect:
Emphasize Topical Relevance
Think of generative AI search as the next evolution of voice search. LLM models weigh a user’s search intent by semantic relevance. Create content that answers conversational search queries and build out your topics in natural and informative ways to optimize for generative AI search.
Creating useful, relevant content that offers real value, context, and information to your user will help your website’s content outshine low-quality, spun content.
Prioritize the Creation of Personalized Content
Google SGE’s ability to offer personalized results may make user segmentation, personalized content, and contextualsearch even more important after its release. Plan to create customized content to help answer and solve for different questions and needs.
Redefine Meaningful Metrics & KPIs
Web traffic may fall if AI-generated answers are thorough enough to impact click-through rates in search, but other metrics may better help marketers understand how their content is being digested and used, such as engagement, attribution, user paths, shares, sentiment, and conversions.
Focus on User Engagement and Retention
Websites may have to focus more time and energy on creating engaging and interactivecontent that offers an even better user experience and more personalization than AI-generated answers. Additionally, certain types of content, such as videos, may also be featured in this new search format more often. As such, diversifying content formats may help content creators gain more visibility and traction online.
Double-Down on Thought Leadership
AI is great at summarizing data that’s already been created, but it can’t create new insights on its own just yet. It can create a replica of what that answer should look like to its users, but ingenuity and new insights are still up to us.
FAQs About Google SGE
Here are some of the most common questions about AI search we hear from clients:
Is Google SGE Available to the Public?
No, Google SGE isn’t available to the public yet. Currently, it’s only available to select users who sign up for beta testing on Google Labs. This will allow more testing and user feedback before the updates are rolled out as a core algorithm update.
When Will I See the Impact From Generative AI Search?
Google plans to initially release Google SGE to one million users exclusively in the United States, gradually increasing to 30 million users by the end of the year. Marketers won’t start to see the full impact of Google’s generative AI search engine until 2024.
How Reliable Is Generative AI Search?
As always, it’s important to remember that while AI can be a hugely useful tool for helping us digest and understand a wide range of complex subjects, that doesn’t mean the results it gives will always be accurate.
AI’s ability to create relevant answers instantaneously seems almost magical, but the reality is just math: LLMs are trained on billions of data sources. AI models map topics, and then make statistical guesses about the words and phrases that should come next. However, that doesn’t mean the answer is correct by default.
On top of predictive errors, there’s also the problem of AI hallucinations, where generative AI creates or cites a false source that doesn’t actually exist to explain its errors. Beta testing in Google Labs seeks to curb these sorts of results and impose safety protocols for users before its large-scale release.
AI tools can only produce answers from datasets they’ve been given and trained on. At scale, data quality assurance becomes harder to control. It’s always a good rule of thumb to fact check anything you read on the internet against reliable sources.
While many of its features are incredible advances in user experience and increased relevance, there are still some growing pains ahead as we collectively learn how to coexist with AI. The AI revolution is experimental territory for Google and users alike, and you can expect Google SGE features to ebb and flow as they figure out how to balance traditional search marketing functionality with AI safety and training protocols.
Will Generative AI Search Replace Traditional Search Engines?
It’s unlikely that generative AI search will replace traditional search engines entirely. As Google has stated, generative AI search results won’t be appropriate for all search queries and topics, such as medical information, financial information, voting information, and other sensitive topics. For these types of topics, search queries will populate similarly to how they currently do.
In the interest of protecting its AI models from bias and toxicity, it will also avoid giving AI-generated results for search terms with value qualifiers.
Will Google Ads Still Have a Place in AI Search?
As their largest revenue driver, Google will continue to display Google Ads in its search results. Search queries that may lead to a transaction will still feature ads within their search results. Google SGE will display ads above the generative AI answer, and throughout search results, much as it does now.
How Will Original Content Creators Be Credited for Their Work?
When Bard first launched, many cited concern that the chatbot’s answers did not cite sources for its answers. Many content creators feared losing visibility and traffic for their work.
Google’s solution? In their SGE reveal, three featured snippets appear next to the AI-generated answer as their answer to the citation concerns. Supposedly, these three websites are featured because those results match the AI-generated answer most closely. Users can also find more sources by clicking the button to the upper right to see more citations.
While traditional search results can be found underneath the answer, the impact on visibility and clickthrough rates for organic search results remains uncertain.
As Google experiments with the ethical and financial ramifications of this new format in Google Labs, content creators have some time to brace for the changes ahead. Optimizing content for SEO will become even more vital, as will building alternative revenue streams to account for potential site traffic losses.
Regardless, Google will need to find a way to reward and credit content creators to continue to build and innovate, or individuals or businesses may suffer unintended consequences for the novelty and user experience of AI-generated answers. After all, Google’s generative AI search only has value if it continues to deliver quality content (that it trains and samples from original content creators).
Is Google SGE Built on Google’s Bard?
No, Google SGE uses different technologies than Google’s Bard. Bard is a chatbot that uses PaLM2 and Google’s very own LLM named Large Model for Dialogue Applications (LaMDA) and trained on a dataset called Infiniset.
Google SGE is a generative AI search technology that uses MUM and PaLM2, and will be trained on Google’s internet indexing data. While it will answer search queries in a conversational way, it will function primarily to deliver relevant search results and concise answers.
Though Google SGE is still in beta, one thing is very clear: this new search experience has the potential to completely revolutionize the ways in which audiences discover and engage with brands. In turn, these changes to users’ search behaviors may end up having a significant impact on the volumes of traffic and conversions that brands are able to drive from both paid and organic search as well. As such, marketers will need to keep a close eye on how Google SGE — and AI as a whole — continues to evolve so that they can adapt their strategies to keep pace with this shift toward more dynamic and personalized digital experiences.
To help make it easier for you to keep up with all these changes, we’re developing a content series focused on exploring the topic of generative AI from a number of different angles. Topics we’re specifically focused on studying and researching include:
Discrimination & bias
Legal challenges associated with AI technology
So, if AI is something you’ve been wanting to learn more about, make sure to check back in regularly to find all the latest research findings and POVs on these topics (and more) from A&G’s department leads.
Want to learn more about what the rollout of Google SGE could mean for your business? Contact us today to schedule a quick 15-minute chat with our Integrated Search Engine Marketing team.
About the Author Olivia Mungal is an SEO strategist at A&G, focused on the always-changing search landscape and its effects on both users and brands.
This article was edited by Katie Bonadies, SEO and Content Strategy lead at A&G.