This blog was originally published on February 8, 2023 and was updated on June 12, 2026.
Back in February 2023, I wrote this blog arguing that ChatGPT wouldn't kill Google. Looking at things three years later, I was right — and I was also wrong in ways I didn't anticipate.
What I got right: Google is still the top search engine.
What I got wrong: I underestimated how quickly even traditional search would change, and how fast we'd move toward using AI-powered results as the primary experience rather than a feature bolted on the side.
No doubt about it, ChatGPT was one of the biggest web innovations we’d seen in years when it launched. It showed the general public the potential of AI assistance like nothing had before. The possibilities seemed endless. Even the creator of Gmail went so far as to say it would “eliminate” the SERP, “destroying the most valuable part of [Google’s] business.”
It turns out he wasn't too far off. He just got the mechanism and the timeline wrong. ChatGPT didn't change the SERP — Google did by introducing AI functionality.
The first-mover disadvantage means emerging, original tech products are destined to become features on established platforms.
There are plenty of times when being the first to market didn’t pay off in the long run.
When Snapchat started to catch on, Instagram (and subsequently Facebook) introduced Stories as a response.
And the pattern is playing out again today with AI search. Perplexity built a genuinely impressive AI-native search experience, and now every other major platform (from Google to Microsoft to Apple) is racing to catch up.
The "code red" moment that occurred in 2022 for Google and other companies
It seemed like Google was caught a little flat-footed when ChatGPT first came out. There were reports of a management-issued "code red" in December 2022, signaling that company leadership saw this technology as a viable threat.
What followed was a cascade of responses from other tech giants: Google announced Bard (now rebranded as Gemini), Microsoft announced Copilot, and Baidu announced its own AI chatbot.
I was right about one key point: Google was uniquely positioned to pick up what OpenAI had started and bring it to the masses.
[INSIGHT] ChatGPT -- and the public’s response to it -- was the best free user research Google could have ever asked for. Three years later, AI Overviews and AI Mode are the result of Google seeing millions of people interact with ChatGPT and understanding the shift toward searching in a different way.
How Google's feedback loops and consistent traffic became their moat
Reinforcement learning saves the day (or rather, the search engine)
Google uses reinforcement learning to train their search algorithms. It’s a machine learning technique where algorithms get rewarded (or punished) to learn and improve. The more the algorithm is rewarded with a desired behavior, the better it reinforces whatever action led to that reward.
In Google’s case, that reward might be a metric like time spent interacting with the search engine, incentivizing its algorithms to show people results that keep them engaged.
You fine-tune these networks through feedback from humans. And the only way to get that feedback is to get the AI in the hands of more humans.
Enter Google: still the most-visited website in the world, with a feedback loop no competitor can replicate at scale.
Google stayed on top, but the story got complicated
Yes, Google has remained the top search engine. But that reason is likely less about Google delivering an objectively better search experience and more about deeply entrenched user habits.
The search experience itself has become fragmented — AI Overviews, AI Mode, traditional results, and ads are layered on top of each other in clunky ways.
The company still spends hundreds of billions of dollars annually to maintain a live map of the internet, and that index advantage is real.
AI search features are transforming how brands interact with consumers
AI Overviews: my prediction, come to fruition
In the original version of this post, I speculated that the future SERP might look "more like a conversation with cited sources." AI Overviews are exactly that: conversational answers, generated by Gemini, delivered directly on the results page with cited sources.
What do AIOs mean for brands and content marketers? The Seer team has been watching closely and our data tells an interesting story: click-through rates declined steadily throughout 2025 as AI Overviews rolled out, but have started to creep back up in 2026.
The search experience may be normalizing, with some users choosing to engage with AI-generated answers and others scrolling past them entirely.
The impact on CTR also varies significantly by query type. Informational queries (i.e. comparison questions, review-style searches, and "how does X work") are now largely answered in the Overview itself. That shows up clearly in the data as lower click-through for those query types.
The opportunity that remains is content with a genuine point of view. Subjective expertise — content that AI can surface but can't replace.
The way people search is changing
Beyond AIOs, Google has launched AI Mode, a fully chat-based search experience. They've also redesigned the Google search bar to actively nudge users toward AI Mode. The "chat-based search engine" that I originally hypothesized about is now the default prompt.
Research from Clickstream Solutions suggests that search behavior now looks more like comparison shopping than a linear "find a result, move forward" experience. People back-scroll, revisit, and cross-reference. It's messier, and harder to attribute.
The real story isn't that ChatGPT kills Google (or vice versa)
Here's what I got wrong in the original framing: I treated it as a binary win-lose scenario where ChatGPT kills Google or it doesn't.
The 2026 reality is that people use both, often simultaneously. They might start a research journey in ChatGPT, pivot to Google for something specific, then head to Perplexity for a comparison. Rather than a situation where platforms are killing each other off, they're becoming different tools for different pieces of the same search journey.
For brands, that means the question isn't "should I optimize for Google or AI search?" Instead, you should be asking "who am I trying to reach, what problem are they trying to solve, and what does genuinely useful content look like for each moment in that journey?"
Now we have GEO (SEO's sibling) to factor into our marketing strategy
SEO isn't dead, but it has a sibling now. GEO (Generative Engine Optimization) is the practice of understanding how AI-powered search platforms decide what to surface, cite, and recommend. It's a new and evolving space that we're still learning about.
What we do know is that strong SEO fundamentals still matter. That means focusing on:
- Clear content that articulates who you are and how you help people
- Consistency in brand positioning
- Content freshness
- Being aware of how other people talk about your brand, because those narratives feed AI models
Sure, Google's AI Mode surfaces different content than ChatGPT does. But the foundation remains the same: be genuinely useful, be clearly findable, and be consistently who you say you are.
Does E-E-A-T still apply?
Search engines will always aim to answer objective queries, and AI has gotten good at answering the informational ones. But that still leaves opportunity for brands with a point of view to create subjective content that showcases real expertise and perspective.
E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) becomes even more important in an AI-powered search world. E-E-A-T is how you get cited by AI, not ranking high on SERP.
So, what does this mean for the future of search?
I was always confident that Google would survive. What surprised me is how quickly "surviving" led to a complete transformation of the search engine (and of search itself).
The thing I'm watching most closely now is wearables. What happens when AI starts searching on our behalf before we even type a query? And how might you influence what a device surfaces before any searches are initiated?
I think applications of AI in the next few years will look different in ways we can't even wrap our heads around now. I expect AI technologies to become more and more integrated into our lives. Apple's moves here will be worth watching.
In the meantime, my advice is to stay curious, keep testing, and keep an eye on how these platforms evolve. The search landscape in 2026 is far more interesting and complex than it was when I first hit publish on this blog.
Stay in tune with Search in an AI world as it continues to evolve by signing up for our newsletter. Insights delivered straight to your inbox.
.png)
-1.png)