Google changed everything on May 14, 2024. If you were watching the Google I/O keynote, you saw Liz Reid, the Head of Search, basically flip the switch on a project that had been simmering in "Labs" for a year. They called it AI Overviews. Most of us knew it as SGE, or Search Generative Experience, but the rebrand signaled something permanent. It wasn't an experiment anymore. It was the new face of the internet.
The announcement wasn't just about a new UI element. It was a fundamental shift in how Google processes information. Instead of just pointing you to a website, Google started trying to be the website. Using a customized Gemini model, the engine began synthesizing data from across the web to answer "ten questions in one." Sounds great for users, right? Maybe. But for anyone running a website, it felt like the floor had just dropped out.
The Day the Search Engine Became an Answer Engine
The blog post that dropped alongside the announcement was titled "Generative AI in Search: Let Google do the searching for you." That headline alone sent shivers down the spines of SEOs everywhere. Basically, Google told the world that for complex queries—stuff like "how to clean a vintage leather couch without ruining the patina"—you wouldn't need to click three different blogs anymore.
Google's AI would just read those blogs for you and spit out a summary.
One of the wildest things about the May 2024 rollout was the scale. Google didn't just trickle this out to a few power users. They launched it to hundreds of millions of people in the U.S. immediately, with a goal of hitting over a billion by the end of the year. It wasn't a soft launch. It was a "we are doing this now" moment.
🔗 Read more: iPhone X screen replacement cost: Why it is still so expensive in 2026
Multi-Step Reasoning and the "Planning" Trap
During the announcement, Google leaned heavily into "multi-step reasoning." This is tech-speak for the AI being able to handle a prompt that has four different requirements. Imagine asking, "Find the best yoga studios in Boston, show me their intro offers, and tell me the walking time from Beacon Hill."
Previously, you’d do three searches. Now, Gemini handles the "legwork."
How it actually works under the hood
Google uses something called Retrieval-Augmented Generation (RAG). It’s not just a chatbot guessing what comes next. It’s actually looking at the live web index, pulling facts from specific pages, and then using the LLM to stitch them together.
But there’s a catch.
Shortly after the May launch, things got... weird. You probably remember the memes. The AI told people to put non-toxic glue on pizza to make the cheese stick. It suggested eating at least one small rock a day for minerals. These weren't "hallucinations" in the traditional sense; the AI was pulling from a satirical Reddit thread or an old joke article and treating it as gospel. Google had to scramble. By May 30, they released a follow-up blog post explaining that they were refining the "guardrails" and limiting the use of user-generated content for health and safety queries.
💡 You might also like: How to make a hologram without losing your mind or your budget
The "Zero-Click" Nightmare for Publishers
Let's talk about the impact on traffic because that's where the real drama is. Early data from places like ZipTie and various SEO agencies suggested that AI Overviews could gobble up 80% of the screen real estate on mobile.
If you're a publisher, this is terrifying.
- Mail Online reported that their click-through rates (CTR) for certain top-ranking terms plummeted.
- The Planet D, a well-known travel blog, saw massive traffic drops because Google was summarizing their itineraries directly.
- Chegg, the education site, saw its stock price tank after the announcement because, well, why pay for a study guide when Google gives you the answer for free?
It’s a bit of a paradox. Google claims that links within the AI Overviews get more clicks than traditional results, but many publishers aren't seeing that reflected in their analytics. Honestly, if the AI tells me exactly how to fix my leaky faucet, I'm probably not clicking the "Learn More" link to read the same thing in 1,000 more words.
What Most People Get Wrong About Ranking Now
A lot of people think that if you rank #1 in organic search, you'll automatically be the source for the AI Overview.
Wrong.
Studies show that only about 12% of the links cited in AI Overviews are the top organic result. In fact, a staggering 45% of the citations come from websites that aren't even in the top 10. This is actually a huge opportunity for smaller, niche sites. If you have a super-specific, high-quality answer to a "long-tail" question, Google might pluck your content and put it right at the top, even if your "domain authority" is lower than a giant like Wikipedia.
👉 See also: American Savings Bank App Explained (Simply)
How to Actually Surivive (and Thrive)
If you're worried about your site disappearing, you need to change your game. The "May 2024" era of search demands a different kind of content.
Stop writing generic "Ultimate Guides."
Google's AI is really good at summarizing generic info. If your article is just a rehash of what's already on the web, the AI will replace you. You need to provide "Information Gain." This is a term Google uses for content that adds something new—original research, a personal experiment, or a unique take that an AI can't just synthesize from other sources.
Actionable Next Steps to Take Right Now
- Audit your "Informational" keywords. Check which of your top-ranking pages are now being buried by an AI Overview. If the traffic is dropping, those pages need more "human" value—add original photos, videos, or first-hand experiences.
- Focus on "E-E-A-T." Experience, Expertise, Authoritativeness, and Trust. Google is leaning on these harder than ever to avoid another "glue on pizza" incident. Make sure your author bios are real and your data is cited.
- Optimize for the "Web" filter. Google quietly added a "Web" tab to the search results in May. It strips away all the AI and ads, showing only blue links. It's a small win for purists, but it's worth monitoring how much of your traffic comes from there.
- Structure for "Snippet-ability." Use clear H2s and H3s that ask a question, and follow them immediately with a concise, 2-3 sentence answer. This makes it easier for the RAG system to "grab" your content for a citation.
- Go Niche or Go Home. Branded searches and highly specific transactional queries are the most resilient. People still want to buy from you, not from a summary of you.
The May 2024 announcement was a "line in the sand" moment. Search isn't just about keywords anymore; it's about being the most reliable, unique source in an ocean of AI-generated noise. It's kinda scary, but for creators who actually know their stuff, there's still plenty of room to win.
To keep your visibility high, start by identifying your "at-risk" content in Search Console and injecting it with proprietary data or unique personal insights that an LLM cannot replicate.
Practical Next Step: Go to your Google Search Console, filter for "Pages," and look for high-impression pages with a declining click-through rate over the last few months. Those are your primary candidates for an "Information Gain" overhaul to win back territory in the AI-driven SERP.