Every few weeks someone asks me whether SEO is dead. The logic is always the same: AI overviews are replacing links. Chatbots can answer everything instantly. Agents will handle our searches. Why bother with SEO anymore?
It’s a fair question. If you’re looking at surface behavior—traffic dropping, click-throughs shrinking—it looks like SEO is collapsing. But when you dig into the data, a different picture emerges.
News publishers are losing search traffic fast.
A SimilarWeb report revealed that total organic search traffic to news sites dropped from over 2.3 billion visits (mid-2024) to less than 1.7 billion (May 2025). That’s a ~26% decline in aggregate traffic. Simultaneously, the percentage of zero-click news searches rose from 56% to 69%.
1. The Numbers Don’t Match the Narrative
Search Engine Journal recently ran an analysis of 25,000 searches. One stat jumped out: websites that rank #1 on Google appear in AI answers about 25% of the time. That’s not “SEO is dead.” That’s SEO bleeding into AI. Ranking in the top 10 is still the biggest predictor of whether your site shows up in ChatGPT or Google’s AI overviews.
Search Engine Land has another datapoint: despite the hype, AI search is still tiny — around 3% of overall search traffic. Two-thirds of consumers say they believe AI will replace search in 5 years, but today, traditional Google still dominates behavior.
At the same time, Writesonic reports that 58% of Google searches now end in zero clicks. That’s the real story. It’s not that AI killed SEO — it’s that Google itself already trained users to expect answers without clicking. AI is just a different UI for that trend.
So, yes, organic traffic is down. In some industries, the drop is brutal — 18% to 64%. But if you zoom out, you see a cycle: every major search change (ads above the fold, featured snippets, knowledge panels) eats into clicks. AI is just the latest bite.
2. Visibility vs. Traffic
Here’s where most of the panic comes from: people confuse AI visibility with SEO traffic.
There are now tools selling dashboards of “AI rankings” — basically telling you whether your brand shows up in ChatGPT answers. But showing up in an AI box isn’t the same as landing a customer. It’s closer to billboard advertising: brand presence, not conversion.
If you treat AI visibility as a KPI, you’re going to waste money. If you treat it as a brand channel — something layered on top of real acquisition — it makes more sense.
3. Search Is Still Multi-Dimensional
Not every search is someone asking “what is X?” or “how do I do Y?”
- Navigational searches: people looking for a specific brand or product. AI will still direct them to a site.
- Transactional searches: people looking to buy. AI can’t fulfill that — it will push them into a click.
These are the battlegrounds where SEO continues to matter. If your site isn’t technically sound, structured, and optimized for discovery, AI responses won’t funnel users your way.
4. Technical SEO Becomes Infrastructure
Think about SEO less as growth hacking and more as plumbing. Schema, site speed, crawlability — this stuff isn’t glamorous, but it makes your product legible to machines and humans.
Search Engine Land flagged this: indexable content, schema markup, E-A-T (expertise, authority, trust), author bios — these signals still feed AI models. They’re the connective tissue between your work and the engines that surface it. Ignore them, and AI just won’t see you.
5. Product-Led SEO Beats Content Farms
This is the piece most startups miss. AI can crank out content faster than any team. Competing with it on volume is a race to the bottom.
The only durable way to win is to build product-led SEO surfaces. Think calculators, benchmarks, glossaries, interactive tools — features that naturally attract searches because they solve a problem. This is the shift from “write 100 blog posts” to “bake SEO into the product itself.”
That’s why I liked Eli Schwartz’s framing in Product-Led SEO: good SEO is product design, not keyword stuffing. The rise of AI only strengthens that point.
6. Humans Still Matter
AI can summarize, remix, and stitch together. What it can’t do is originate ideas or inject lived context.
That’s why the best SEO going forward will come from companies that combine machine-readable structure with human-level insights. Put bluntly: AI can’t interview your customers, can’t know the quirks of your industry, and can’t build a product that actually solves a pain. It can only amplify what’s already there.
7. What Founders Should Actually Do
If you’re running a startup right now, the SEO playbook shifts, but it’s not gone:
- Stop chasing AI rankings. Visibility isn’t the same as acquisition.
- Double down on technical SEO. Make your product crawlable and machine-friendly.
- Think product-led SEO. Build surfaces that naturally pull queries (tools, benchmarks, data sets).
- Accept lower organic CTR. Zero-click searches and AI boxes are here to stay. Compensate with owned channels (email, community, brand).
- Invest in trust signals. E-A-T isn’t a buzzword. Author bios, citations, data transparency — these are the assets AI models and search engines reward.
Closing Thought
AI didn’t kill SEO. It just exposed how much of SEO was always about chasing algorithms instead of serving users.
The game now is harder but cleaner: if you’re building real content, structured properly, and tied to a product people want — you’ll show up, whether it’s in a blue link, an AI overview, or the next interface we haven’t seen yet.
For everyone else, the free traffic party is over.