From SGE to AI Overviews: A Complete Timeline, as Told Through Different SEO Perspectives
By Ethan Lazuk
Last updated:
Give this a sec to load. ⏱️ … It has a lot of tweets embedded. 😉
You can’t throw a rock these days without hitting a new story about Google’s AI Overviews (formerly SGE).
Just don’t eat that rock, please. 😅
But with so much AIO news coming out, it’s hard to keep track of (let alone remember) everything.
So I worked back to create a complete SGE-to-AI Overviews timeline (which I’ll keep updated).
This is told largely through the perspectives shared by SEO professionals, publications, and Googlers on social media.
We’ll cover everything from when SGE was announced on May 10th, 2023 (at Google I/O) to after it was rebranded as AI Overviews on May 14th, 2024 (at the following I/O).
This SGE-to-AIO history has 3 phases, which I’ve divided into months:
- “Pre-SGE era”: Just before Search Generative Experience was announced at Google I/O 2023 until its release on May 25th.
- “SGE era”: The year of experimentation with SGE in Google Search Labs.
- “AI Overviews era”: The official roll out of AIO to U.S. users at Google I/O 2024 (May 14th), with some aspects still in Search Labs.
Quick notes:
- I’ll occasionally change embeds to links when I add new tweets to ensure this loads on mobile.
- If a thread is condensed, I’ve added a note below it so you know to expand. (Lots of good stuff in those threads!)
- Lastly, these tweets came mostly from Hamsterdam SEO news recaps. You can visit those to find more info from these periods.
Ok! Let’s begin our journey from SGE to AIO, starting in the pre-SGE era …
The pre-SGE era (before May 2023)
This is an interesting period, not only for considering what news came out about Google before SGE rolled out, but also to see what actually rolled out.
Before we get to the tweets, here’s some historical context for the pre-SGE era:
- Google’s use of language models goes back decades, but earlier systems were more based on statistical analysis than deep neural networks.
- RankBrain (a deep learning model) was called the first AI system in Google Search in 2015, but neural networks were likely part of the equation before that.
- A real NLP breakthrough, in general, came in 2017 with transformers and self-attention, which led to contextual embeddings and models like BERT and GPT.
- ChatGPT became a consumer product in November 2022, and ChatGPT Plus was released in February 2023, changing the public’s perception of AI and introducing LLMs more broadly.
- The New Bing (now called Copilot) also began rolling out in late February 2023.
- Google also announced Bard in February of 2023, which used a LaMDA conversational model, different from today’s Gemini (which was then announced at Google I/O as being “still in training”).
- Bard (now called Gemini) started being released to testers (as an experiment) in March of 2023, and its waitlist was removed on May 10th, 2023 (at Google I/O).
- Google I/O 2023 is also where SGE was introduced.
I started compiling weekly SEO recaps in April of 2023, so that’s where I’ll begin this timeline.
April 2023
This is the month just before SGE was announced at Google I/O 2023. This period comes on the heels of the introductions of Google Bard and the New Bing (Microsoft) and amidst the growing fervor over ChatGPT Plus (OpenAI).
Related tweets to check out:
May 2023
This includes the two-week period between when SGE was announced at Google I/O (May 10th, 2023) and when access started becoming available in Google Search Labs (May 25th).
Related tweets to check out:
- Google I/O event marks a turning point for digital marketing (Duane Forrester on X).
- Amazon should be worried (Mordy Oberstein on X).
- Fernando Maciá article shared on X (very good).
- A major shake-up is coming to Google’s search results (Brodie Clark on X).
- SEO is not dead but could lose relevance as a traffic channel (Olaf Kopp on X).
- Google releases 19-page report on SGE (Chris Long on X).
The SGE era (May 2023 to May 2024)
This section covers the one-year period (or thereabouts) after we all got access to SGE as a Search Labs experiment — that first “E” in E-E-A-T, if you will.
Something I noticed, looking back on this time, is the discussion kind of refocused from the macro (future of SEO) to the micro (how SGE works).
Here’s some context to keep in mind for this era:
- We had both the third helpful content update (September of 2023) and March 2024 core and spam update.
- There was ongoing news about spam on Google in early 2024 (hence that March spam update).
- OpenAI’s GPT-4 rolled out just four days after Google I/O 2023, and GPT-4o rolled out one day before Google I/O 2024. (Not going to lie, I use Gemini daily and ChatGPT regularly, and GPT-4o is pretty awesome, but I need both …)
- The New Bing was rebranded as Copilot in November of 2023. (Bing had other innovations, like Deep Search, as well.)
- Perplexity’s growth (January 2024).
And so, the SGE-era begins!
May 2023
This covers the final week of May 2023, which was the first week after SGE rolled out on the 25th. The value of this section, in my opinion, is seeing how SGE looked or functioned early on, and how it evolved into AI Overviews.
Related tweets to check out:
- A few thoughts on my initial Google SGE experience (Aleyda Solis on X).
- SGE screenshots compared with featured snippets (Nick LeRoy on X).
- Article on “Google’s NEW AI Search Results” (Joe Hall on X).
June 2023
This was the first full month after SGE was available, so there were a lot of discoveries happening, as well as officially announced updates from Google.
The early days of SGE seemed to have more focus on its implications for Search, searchers, and SEO strategies, but this is when the attention shifted more to tracking different types of SGE results and the evolution of its UI.
July 2023
This is now the second month after the SGE rollout, where its novelty seems to have worn off a bit, at least based on the number of tweets I found related to it.
This next tweet got deleted, but I left a note about its context (from Hamsterdam Part 14):
August 2023
This month involved a long-awaited core update from Google, so SGE news was limited. SGE also was expanded beyond the U.S. this month for the first time to India and Japan, leading to more people sharing findings in later months.
September 2023
This was the month of the third helpful content update, aka, the start of a busy time for SEOs. 😉
Related tweets to check out:
October 2023 – May 2024
Coming soon!
But let’s now get to the heart of the matter, the official rollout of AI Overviews (formerly SGE) …
AI Overviews (May 2024 onward)
AI Overviews rolled out at Google I/O on May 14th, 2024, to all U.S. users, with up to a billion more users globally coming by the end of the year.
In recent days (June of 2024), the visibility of AIO seems lower. I was personally struggling to trigger them for many queries, but now seeing them pick back up. (This may have been related to the viral strange AIO answers and Liz Reid’s response post, which may have led to a temporary lull in AIO visibility while updates were fleshed out.)
I would also say it’s early days for tracking AIO data, especially with no GSC data to reference. Though we now have more articles about that, as well.
And so, let’s explore the official Gemini era of Search, when AI Overviews begin!
May 2024
This covers the second half of May, following Google I/O and the rollout of AI Overviews. Some odd responses went viral, leading to updates and a response from Liz Reid.
Related tweets to check out:
- Mini case study: getting a site ranking in multiple AIOs (Cyrus Shepard on X).
- Google explains how it’s improving AI Overviews (Barry Schwartz on X). (After reading Barry’s summary and Liz Reid’s post, I visited the links for additional background.)
June 2024
This is the first month that AI Overviews were rolled out, which included updates, fluctuations in visibility, and more research on possible impacts.
Related tweets to check out:
- Google AI Overviews visibility drops (15% of queries) (SEL on X).
- A new Free Chrome Extension to Analyze AI Overviews (Aleyda Solis on X).
- Google posts AI Overviews FAQs (Barry Schwartz on X).
- Google showing more AIOs for health queries (Barry Schwartz on X).
- Analyses: Google AI Overviews showing less often, with less Reddit (Danny Goodwin via SEL on X).
July 2024
We’re seeing more great studies and articles on AIO’s impact and how it works.
Related tweets to check out:
- Zero click searches in 2024 (with impact of AIO in May) (Rand Fishkin on X).
- Google tests back to AI Overview button (Barry Schwartz on X).
Closing thoughts (human generated) 🤘
I hope you’ve enjoyed this journey through the history from SGE to AI Overviews!
SEO has a weird sense of history, though, doesn’t it?
I mean, we have basic concepts — technical SEO, helpful content, and authoritativeness — which are pretty timeless.
But the search landscape evolves quickly.
Who would’ve thought, even a few years ago, we’d have the efficiencies LLMs can provide for SEO work, for example?
Part of the reason I believe the field develops so fast, especially nowadays — aside from search engines always working to prevent spammers — is that modern IR systems are tied more to machine learning, DNNs, and LLMs.
I refer to our current search era as the “post-HCU world,” where SEO strategies must go beyond keywords and focus on semantic concepts and user-centric approaches — and even culture 😉 — in support of brand awareness and business initiatives.
Studying neural networks can help improve our comprehension of how today’s search ranking systems function — or help explain associated “rankings volatility” — even still, these ML-based systems are likely detecting semantic relevance, authority, and patterns in user interaction data (and making predictions from that) that we ourselves may never see or comprehend.
Whether or not AI Overviews becomes a more prominent and permanent fixture in Search or is merely a temporary interface on the way to something else (like more chatbot and AI agent usage), by reviewing SGE’s development and placing AIOs in a wider historical context, we can appreciate why the future of SEO depends less on individual tactics and more on holistic and brand-focused strategies.
Outro
Stay tuned for future updates and improvements to this article.
Feel free to comment with thoughts or questions (or contact me) or view related content below.
Until next time, enjoy the vibes:
Thanks for reading. Happy optimizing! 🙂
Leave a Reply