Yes. Keywords still matter in SEO, but their job has changed. They used to be the target. Now they’re the best clue you have about what people need, how they describe it, and what they expect to find when they search.
That shift matters even more as AI search grows. Google, ChatGPT and other AI tools read a question for its meaning, then pull answers from sources that explain the topic well. A page built to repeat a phrase rarely makes that cut. A page built to answer the need behind the phrase often does.
The short answer for busy teams.
Keywords still earn their place in four ways:
- Research. They show the questions, problems and language your audience uses.
- Intent. The results for a keyword tell you what searchers expect to see.
- Structure. Grouped keywords shape your topics, pages and internal links.
- Signals. Clear page titles, headings and copy still help search engines and AI tools understand a page.
Keywords hold teams back when they’re used as a density target, a ranking scoreboard or the reason a page exists.
Google moved from matching words to understanding meaning.
The change happened in steps over more than a decade:
- 2012, Penguin: lowered rankings for sites that used manipulative tactics like keyword stuffing.
- 2013, Hummingbird: introduced semantic understanding, so the meaning of a query mattered more than its exact words.
- 2015, RankBrain: added machine learning to interpret queries Google hadn’t seen before.
- 2019, BERT: improved how Google reads natural language and the intent behind it.
- 2022 to 2024, helpful content: rewarded content written for people over content written for search engines. In March 2024, Google folded this into its core ranking systems.
- 2024 to 2025, AI Overviews and AI Mode: Google began answering many searches with AI-written summaries that cite a handful of sources.
None of these updates removed keywords. Each one made the meaning behind a search more important than the exact words in it.
Why so much of the SEO industry stopped talking about keywords.
Editors, practitioners and leading SEO publications now talk far less about keywords and far more about topics, intent, entities and the prompts people type into AI tools. That shift is real. A prompt in ChatGPT or Google AI Mode is often a full sentence or more, and no two people phrase it the same way, so there’s no single phrase to target.
Keywords didn’t disappear in that shift. They moved from the center of the strategy to the research behind it. They’re still the clearest record of how people describe what they need, and search tools still report demand by keyword. What changed is the unit of planning. Teams now plan around a topic and the questions people ask about it, not a phrase.
Writing to hit a keyword is outdated. Researching keywords is not.
We often hear that “writing for SEO” is dead. That’s partly right. Writing around a keyword list, repeating a phrase a set number of times, and judging content by its keyword density stopped working years ago.
Keyword research is a different thing. It’s still one of the fastest ways to learn how your buyers think. A hundred people might ask the same question a hundred different ways, and the patterns in that language tell you what to explain, what to compare and what to answer first.
Should you still optimize a page for one specific keyword?
Sometimes. A narrow focus makes sense when one term carries real business value and has a clear, consistent intent. A core service page, a local service page or a product category often fits that description.
A single-keyword focus becomes a problem when it stands in for strategy. Chasing one high-volume phrase can pull a page away from what its buyers need. It can also hide the dozens of longer, more specific searches that bring in better leads. Paid search teams have had a version of this debate for years with single keyword ad groups, and the lesson carries over. Tight focus helps when it serves the searcher, and hurts when it only serves a report.
Our rule of thumb is to pick a primary topic for every page, support it with the related terms people use, and let the content answer the full need.
Start with strategy, then plan around topics.
We recommend putting strategy before tactics. Start with the business goals and the buyer’s questions, then use keyword data to confirm the language, size the opportunity and spot gaps. Group terms into topics, check the live search results to confirm intent, and map each topic to one page so pages don’t compete with each other. Treat keyword data less as a list of targets and more as a way to understand demand.
Rankings alone no longer tell you if SEO is working.
Keyword rankings are harder to read than they used to be. Search Console groups or hides many long-tail queries. AI Overviews and AI tools answer questions without a click. Voice and visual search send little keyword data at all.
That’s why we look at a wider set of measures:
- Leads and other conversions from organic search, and their quality
- Visibility across groups of related queries, not single terms
- Mentions and citations in AI answers
- Engagement on the pages that matter to the business
Keywords in AI search work the same way, only more so.
AI tools read a question as a whole and look for sources that explain the topic clearly and credibly. Clear structure, direct answers, real expertise and consistent information about your brand across the web all help. We cover the bigger picture in SEO vs GEO vs AEO vs LLMO and on our AI search page.
The bottom line.
Keywords still matter in SEO. Use them to understand your audience, confirm intent and organize your content, then write for the person on the other side of the search. If you’d like help putting that into practice, see how our SEO services work.