Join Bastian as he delves into Google's AI efforts and how the AI Overviews (formerly known as Search Generation Experience - SGE) will impact your eCommerce or content publishing business.
give an immediate sense of what generative AI is currently being used for: Source: https://pa.ag/4brddtI - Technical Assistance & Troubleshooting (23%) - Content Creation & Editing (22%) - Personal & Professional Support (17%) - Learning & Education (15%) - Creativity & Recreation (13%) - Research, Analysis & Decision Making (10%)
to regulate AI The Artificial Intelligence Act (AIA) is a legal framework for trustworthy artificial intelligence in Europe. Source: https://pa.ag/3ufz2wg The AIA pursues a risk-based approach for AI with a European scope. It is primarily concerned with: ▪ The regulation of prohibited practices ▪ Requirements for high-risk AI ▪ Transparency and accountability
by 2025... Amazon, for example, invested $4 billion in Anthropic, Google is adding another $2 billion to the pot... And these are just two of the larger deals in recent months. Source: https://pa.ag/47nSJjv
Google Google's CEO Sundar Pichai summarised their I/O 2016 conference, stating: “It’s an important shift from a mobile first world to an AI first world" Source: https://pa.ag/3GgHiip In 2016, a few months after becoming CEO of Google, Sundar Pichai made a sweeping proclamation: Google, whose name had become synonymous with search, would now be an “AI-first” company.
led to Google launching two completely unfinished products one after the other. Sources: https://pa.ag/3SL0o7L & https://pa.ag/46ckFpQ & https://pa.ag/3uct7bl
transforming the Search experience with generative AI. When using SGE, people will notice their search results page with familiar web results, organized in a new way […]”
click AI-generated textual summary of the topic (comes in various formats) Entry into conversation mode, incl. AI-generated suggestions/query refinements Thematically appropriate image, usually taken from the first (old) organic result Three (or more) organic results; link directly externally. Main sources for the snippet content
moving forward? Google is considering charging for new “premium” features powered by Gen AI. This would be the biggest ever shake-up of its search business so far: Source: https://pa.ag/3xnIgbc
2024? Apart from the official renaming of SGE to AI Overviews, not that much, really – except for an actual public roll-out! US rollout first, (starting) this week AI overviews will be available to all US searchers on Google starting this week, which means hundreds of millions of searchers. No additional rollout dates AI Overviews will launch in more countries soon (but no dates are given so far) and be available to more than a billion users by the end of the year. No detailed data in GSC AI Overviews are tracked in GSC with impression, clicks & average position, but won’t be shown separately (like, e.g., Featured Snippets) More features in Search Labs Google is rolling out new AI features, including multistep reasoning & planning capabilities, AI-organised SERPs, etc.
“Soon, you’ll be able to adjust your AI Overview with options to simplify the language or break it down in more detail. This can be particularly useful if you’re new to a topic […]” [Liz Reid, Google] Source: https://pa.ag/4bs7lRB
greatly Current median: 1,050px - but in some cases they’re enormous and more than 4,500px in height: Zoom@33%; 4,580px Zoom@50%; 2,500px Zoom@50%; 1,800px
an SGE answer Basis: 100,000 keywords in mid. February 2024 (left chart) vs. 91,000 keywords in mid. October 2023 (right chart) where around 40% of test queries (promptly) provided an SGE snippet: Sources: https://pa.ag/49ku6F3 & https://pa.ag/3QzDiyd Presence of SGE snippet 39,66% 30,17% 30,17% Has_Snapshot No_Generate_Button No_Snapshot
to stay! In Google‘s own words: "User feedback has been very positive so far." With AI Overviews, people are visiting a greater diversity of websites for help with more complex questions. And we see that the links included in AI Overviews get more clicks than if the page had appeared as a traditional web listing for that query […] Sundar Pichai, CEO Google & Alphabet
result (for users) is preferable to clicking on their own website has had problems long before Google started testing SGE... Commodity content has long been doomed to fail
summaries fundamentally change the business that Google was built on ("give us content, we'll give you traffic") by taking even more content from the source and showing it directly - keeping more of the search journey on Google.
is a coincidence? "Currently available in the contiguous United States […] MetNet-3 is helping bring accurate and reliable weather information to people in multiple countries and languages.” Source: https://pa.ag/3uaeqVX Weather integration in AI Overviews
eCommerce SERPs... outside of the shopping tab! ▪ Price history in the SERPs ▪ The filter bar is switched on by default, showing significantly more popular products ▪ On the other hand, the products pull more content directly from the websites ▪ The buy now button immediately starts the checkout process
Google Shopping From “Virtually try it on“ to “Bring your vision to life with image gen” (shopping for clothing styles similar to whatever you had in mind) and “More from your favorite brands” – Googles has introduced lots of new shopping functions! Source: https://pa.ag/3TLgbSJ
https://pa.ag/3amVV3L & https://pa.ag/46isw5g My team and I spent a great deal of time providing high-quality answers to questions that haven’t even been asked yet. Prabhakar Raghavan (SVP, Google) Reading recommendation
by combining (multiple) queries, user profiles and ranking data: Source: https://pa.ag/3u8gLRs It’s a little bit more time to really go deep in understanding the query […] this is all about being able to have an orchestration between the queries itself and the ranking and the profile of a user to really go deeper in understanding what the user is looking for. Fabrice Chanel (Principal PM, Bing)
Latent spaces give a machine a condensed understanding of the world - through spatial representation. And the model learns from this automatically! Source: https://pa.ag/3sBne6X People can understand topics by encoding observed events in the brain in compressed form. For example, we can recognise a dog on the street without remembering every detail of its appearance.