The End of SEO As We Know It

This Week

This week, it’s all about search.

Not the kind that gives you ten blue links. The kind where AI does the work for you — pulling research, writing summaries, and deciding what content is worth showing in the first place.

We’re talking AI search, SEO shifts, and what happens when visibility moves from the top of Google to inside your chatbot’s answer.

Whether you're creating content, optimizing for organic, or just trying to keep up — this is the issue to read.

Let’s get into it.

What Happens When AI Stop Searching and Starts Thinking?

AI is shifting from answering questions to doing research.

Ethan Mollick, AI expert and professor at Wharton, believes LLMs are now capable of something bigger: producing research with depth, nuance, and speed that rivals human experts. 

In his article The End of Search, The Beginning of Research, he lays out where this is heading and why it matters.

Here are three big shifts he highlights:

1. AI can now “think” before answering. With reasoning models (like OpenAI’s Reasoner), AI processes information first, improving its ability to solve complex problems. The longer it thinks, the better the output.
   

2. Narrow AI agents are already outperforming expectations. Tools like OpenAI’s Deep Research can conduct high-level academic research, resolve contradictions in sources, and synthesize findings into graduate-level reports — all within minutes.
   

3. Humans are moving from researchers to reviewers. These AI systems can complete work that used to take teams of skilled professionals. The human role is shifting toward guiding, validating, and refining AI output, rather than starting from scratch.

OpenAI, Perplexity, and Google have all released Deep Research features that take this even further. These tools search wide-ranging sources and generate 8–10 page reports with citations. It’s a major shift in how we interact with information and how knowledge work will get done from here on out.

If AI tools can search, synthesize, and deliver content on demand, traditional search traffic is going to drop, especially for top-of-funnel and long-tail queries. Content needs to shift from ranking-focused to source-worthy. 

That means prioritizing original thinking, research, and depth that AI will pull from — even if fewer humans hit your site. Visibility is moving from search results to influence inside the AI’s output.

Mollick has a ton of great advice on using AI, so if you haven’t already subscribed to his blog you should head over there now: One Useful Thing.

AI Overviews Impact on SEO

Kevin Indig — one of the sharpest minds in SEO and AI search — just did what most of us wish we had time to do. He analyzed 19 different studies on Google’s AI Overviews to figure out exactly how they’re affecting search, clicks, and traffic.

I went through his full breakdown (so you don’t have to) and here’s what stood out.

First: How often do AI Overviews show up? 

It’s less than you think, but gaining steam. Depending on the industry, AI Overviews show up in up to 47% of searches. The biggest areas? Health, shopping, and “how-to” queries.

Second: They’re crushing click-through rates.

When AI Overviews show up, organic results get pushed down by 900+ pixels. That’s basically “goodbye” to first-page visibility on mobile. Most studies report up to a 50% drop in CTR for traditional results.

Third: Ranking in AI Overviews ≠ Ranking in SERPs.

Just because you rank #1 doesn’t mean you’re in the AI Overview. Only 20-40% of citations come from top-ranking results. AI Overviews pulls sources that sound similar to its answer, not just those with keywords. This means text similarity (not ranking) should now be part of your SEO strategy.

Fourth: There’s limited room to win.

On average, AI Overviews cite 5-8 sources, and 99% of sources are cited once. That means you’re fighting for tight real estate. If you’re not cited, your content’s likely invisible.

So what does Kevin recommend you do?

• Focus on long-tail, informational keywords — that’s where most AI Overviews live.

• Optimize for clear, concise answers — think 90-170 words, similar to Featured Snippets.

• Test YouTube and LinkedIn content — these domains are heavily cited.

• Check your Knowledge Graph presence — AI Overviews often lean on it for facts.

The takeaway? Being cited in AI Overviews is just as important as ranking in Google. And SEO is moving from focusing on the top result to being visibility inside AI answers.

You can find Kevin’s write up at his blog: Growth Memo.

Google’s New AI Mode is Taking Over Search

Google just rolled out AI Mode, an experimental feature powered by Gemini 2.0, and it’s a glimpse into the future of search.

With AI Mode, search feels more like a conversation than a hunt. You ask one question, then another, and AI Mode keeps rolling out answers. It also breaks down complex queries, compares options, pulls in visuals, and thinks through answers using data. 

I asked it a range of questions on a complicated topic (NDC for any of my travel friends 😄) and here’s what I noticed:

• The results are more answer than link driven, but sources are still available

• The follow-up flow feels natural — you can dig deeper into your search without starting over

• It builds out comparison tables when analyzing multiple items or solutions

• Unlike Google’s AI Overviews, AI Mode is available for all search queries

Over a billion people are now using Google’s AI Overviews, and AI Mode is expanding fast. 

If you’d like to test it out yourself, sign up at Google Labs and you can get all the details at the Google blog. 

ChatGPT Just Saved me $3,000 in SEO Analysis

I’ve been testing ChatGPT’s Deep Research feature, and decided to throw it a real challenge: a full SEO audit for a client’s site.

Instead of jumping straight into a response, it paused and asked a few sharp questions—what tools I use, who the competitors are, the target audience. Then it got to work.

Fifteen minutes later, I had a clean, detailed, 20-page report of very specific feedback. It ran speed tests, flagged broken links and rewrote metadata.

Some of the better recommendations it surfaced:

• Tighten up CTAs on high-traffic pages

• Remove jargon that’s likely confusing new users

• Refresh older blog content to stay competitive

I sent it to a couple SEO friends to double-check. They said it was a solid starting draft with some rough edges, but workable.

One big gap? No mention of AI search.

It missed things like optimizing for AI Overviews or ChatGPT’s browsing features. But when I followed up and asked how to adapt it for generative search, it quickly added a solid section covering that too.

Would I use the full report as a final deliverable? No. But for 15 minutes of effort, it was a killer draft.

Want to run the same audit for your website? I’ll be sharing the exact prompt I used later this week on LinkedIn, so be sure to follow me there if you want to grab it.