For about fifteen years, keyword strategy was the closest thing SEO had to a science. Find the terms people search. Understand the intent behind them. Build content that answers the question better than the current results. Rank. Get traffic. Repeat.
This is not how it works anymore.
Google’s AI Overviews answer the question before the user clicks anything. For informational queries — the “how,” “what,” “why” questions that formed the backbone of content marketing strategy — a significant percentage of users now get their answer without ever visiting a website.
The question isn’t whether this is happening. The question is: what fills the void?
What’s Actually Getting Killed
Before we talk about what survives, let’s be specific about what doesn’t.
Definition and explanation content is nearly dead as a traffic source. “What is content marketing?” “How does compound interest work?” “What’s the difference between X and Y?” If an AI can accurately answer it, and most of the time it can, you’re not getting the click.
Listicles without genuine curation are in serious trouble. “10 best CRMs” worked when it required a human to compile and compare. It works less well when an AI can synthesize that comparison in real time.
Thin how-to content — the kind that walks you through a process Google could now demonstrate in a sidebar — is losing.
What’s not getting killed: anything that requires genuine expertise, recent events, original research, personal experience, or a perspective that someone would specifically seek out.
The Entity Shift
The underlying structural change is from keyword-based to entity-based search.
Google increasingly understands content in terms of entities — people, organizations, concepts, places — and the relationships between them. When you’re a recognized entity with a clear area of expertise, your content surfaces differently than when you’re just a domain with keywords.
This is why personal brands and organizations with genuine authority in a specific domain are faring better than sites that were built primarily as keyword targeting vehicles.
Practical implication: The question shifts from “what are people searching for?” to “what is this site or person the definitive source on?” These are different strategic questions and they lead to different content choices.
What Actually Works Now
Original data and research. AI Overviews pull from existing content. They can’t synthesize original research that doesn’t exist yet. If you’ve run a survey, analyzed a dataset, or synthesized proprietary industry data, you have something that cannot be replicated by an AI answer.
Experience-led content. “I tried this for six months and here’s what I found” is hard to replicate because it actually requires having done the thing. The more specific and creditable your first-person evidence, the harder it is to replicate.
Contrarian takes with receipts. Content that challenges the conventional wisdom in a defensible way drives the kind of engagement — comments, shares, links, newsletter forwards — that signals genuine authority. A good controversial argument with solid support does things a keyword article never could.
Community and audience ownership. Email lists, Slack groups, Discord servers, podcast audiences. These bypass search entirely. The search traffic you’re losing is the traffic you never really owned anyway.
Bottom-funnel specificity. Product comparisons, integrations, pricing pages, use-case specific content — these queries still drive clicks because the decision-making complexity means people want to explore, not just get an answer.
The Honest Reckoning
Here’s the difficult truth: a lot of content that ranked well for the last decade was ranking because it was the best available answer in a pre-AI system, not because it was genuinely valuable.
The AI filter is, in a sense, a quality filter. Content that existed to capture keyword traffic and redirect it somewhere else is the content that’s disappearing. Content that exists because a specific person with specific knowledge wrote something worth reading is holding up better.
The pivot isn’t purely strategic — it’s about actually having something worth saying.
For brands that have been playing the keyword volume game, this is a painful reckoning. For the ones who’ve been building genuine expertise and audience trust, it looks more like an opportunity.