The Question Everyone Is Asking Wrong
"Will ChatGPT kill SEO?"
That's the question that's been circulating since OpenAI released ChatGPT in late 2022 and search volume dropped measurably for certain informational queries. The question is understandable. The framing is wrong.
The right question is: how has the existence of LLMs changed what search engines need to reward, and what does that mean for the content strategies that work?
That question has a more interesting — and more actionable — answer.
What Actually Changed
When ChatGPT launched, something real happened to search behavior. For a specific category of queries — the kind where someone needs a quick, synthesized answer to a factual or procedural question — LLMs became a meaningful alternative to Google. "How do I write a Python function that does X?" "What's the difference between a Roth IRA and a traditional IRA?" "Explain quantum entanglement in plain English."
These are synthesis queries. The user wants information assembled and explained, not a list of links to go explore.
Google's response was predictable and significant: they built AI Overviews (formerly SGE) directly into their search results, attempting to keep users on Google while providing LLM-style synthesized answers. For synthesis queries, Google now answers in the results — and the sites that used to rank for those queries have seen traffic drops ranging from significant to catastrophic.
This is the real change. Not that ChatGPT killed SEO — but that Google itself fundamentally changed how it handles a whole category of queries in response to LLM competition.
What Didn't Change (And People Miss This)
For all the noise about AI killing search, a large majority of commercial search intent is completely unaffected by LLMs. Consider:
- Transactional queries: "buy running shoes size 11 wide" — ChatGPT cannot fulfill this
- Local queries: "best Italian restaurant near me" — LLMs don't know where you are or what's open
- Research and comparison: "best CRM for small business 2025" — users want current, opinionated, human-verified comparisons, not synthesized averages
- Navigational queries: Someone looking for a specific site or brand — they're going to Google
- News and current events: LLMs have training cutoffs; search has real-time indexing
- ChatGPT didn't kill SEO — it shifted which queries flow to LLMs vs. search, primarily informational head terms
- Google's AI Overviews are the real disruption — Google is now answering synthesis queries in-SERP, reducing click-through to content
- Transactional, local, specific, and current queries are largely unaffected — if your strategy covers these, you're insulated
- E-E-A-T matters more now — LLMs can synthesize average knowledge; Google needs to surface what LLMs can't produce
- Long-tail, high-specificity queries retain full value — too specific for AI Overview, too commercially important for users to trust LLM answers
- To appear in AI Overviews: rank organically, structure clearly, make claims quotable and specific
The informational query category — how-to articles, definition pieces, general explainers — is where the disruption is concentrated. If your SEO strategy is built primarily on capturing informational traffic through thin, synthesized content, you're in trouble. If it's built on authoritative, commercial, local, or deeply specific expert content, you're largely insulated.
The brands that got hurt worst were the ones whose content strategy was essentially "produce synthesized answers to common questions at scale." That's exactly what LLMs do better than any human content team. The brands that have held are the ones with genuine expert knowledge, real product coverage, or local presence that LLMs can't replicate.
The E-E-A-T Acceleration
Google's response to the LLM era has accelerated something that was already in motion: the premium on first-hand experience, genuine expertise, and real-world authority.
The logic is straightforward. If LLMs can synthesize average knowledge, then average knowledge has no value in search results. Google needs to surface content that LLMs can't produce — content that comes from direct experience, proprietary data, human judgment, and real-world accountability.
This means E-E-A-T signals matter more now than they ever have. Demonstrated personal experience. Named expert authors. Original research. Proprietary case studies. Opinions formed from doing the actual work.
At X Network, we've seen this play out with clients. The articles that are holding rankings — or gaining — in the post-ChatGPT environment are ones with specific, attributable claims. "I tried X approach on Y type of campaign and saw Z result" consistently outperforms "research suggests that X approach generally produces better results."
The former can't be generated from training data. The latter is indistinguishable from AI output.
What This Means for Keyword Strategy
The keyword implications are significant.
Informational head terms — "what is SEO," "how to write a cover letter," "what does a digital marketing agency do" — have declining search traffic and declining value. Google is answering many of these in AI Overviews. The clicks that used to flow to content covering these queries are being absorbed by the SERP itself.
The queries that retain value are:
Long-tail, high-specificity queries: "best SEO agency for SaaS companies with enterprise sales cycles" — these are too specific for LLMs to answer well, too specific for Google to generate a useful AI Overview on, and too commercially valuable for users to trust an LLM answer without verification.
Queries with strong commercial or transactional intent: The path to purchase still runs through search. Someone researching a high-consideration purchase — software, services, large consumer goods — wants credible sources, not a synthesized answer.
Queries requiring current information: Training cutoffs mean LLMs age poorly for anything time-sensitive. Event coverage, product launches, regulation changes, market data — these still need current, indexed content.
Queries requiring local knowledge: Any business with local relevance is protected from LLM disruption in search. The local search ecosystem is still fundamentally Google-driven.
The content strategy that follows from this: less content targeting generic informational queries, more content targeting high-specificity queries with genuine commercial intent or first-person expertise.
How to Position for AI Overviews
AI Overviews appear at the top of Google results for informational queries and synthesize answers from multiple sources. The question for content creators is: how do you get cited?
The pattern I've observed in my own clients' coverage: AI Overviews tend to cite sources that are:
1. Already ranking in the top 3-10 organic positions for the query
2. Structured clearly with defined headings, specific claims, and quotable sentences
3. Authoritative on the topic as measured by existing E-E-A-T signals
4. Directly responsive to the specific question being asked
The practical implication: ranking for AI Overview citations is not a fundamentally different game from ranking in organic results. The same content quality, authority, and technical optimization that produces organic rankings tends to produce AI Overview citations. The addition is structural clarity — content that's easy for Google's LLM to extract and attribute.
The Bigger Picture: Search Is Not Dead
I've watched SEO declared dead approximately a dozen times since I started in this industry. Every major algorithm update. Every new platform. Every shift in how people discover content.
Search isn't dead. It's restructuring. The queries that used to flow to generic content farms are flowing to LLMs instead — and that's probably net positive for the internet. The queries with real commercial value, real specificity, and real need for expert human judgment are holding.
The brands and agencies that are panicking right now are the ones whose strategies were built on the edge of sustainability anyway. They were winning by manufacturing content at volume — answering every question anyone might ask, regardless of whether they had anything genuine to say.
That edge has collapsed. The strategies that remain viable — and the ones that will compound — are the ones built on genuine expertise, first-hand experience, and content that can only exist because a specific person with specific knowledge wrote it.
That's not a worse SEO landscape. That's a better one.