TL;DR. Keyword research starts with the words, phrases, and questions real people actually type, then narrows that list down by volume, difficulty, and intent. 91.8% of all searches are long-tail keywords, and B2B companies that do this well report 702 to 1,389% ROI from SEO, while AI search has added a second, much longer tail on top of the first: AI prompts average 25 or more words, creating a tail roughly four times larger than the one traditional keyword tools were built to measure.
What is keyword research?
Keyword research identifies the specific words, phrases, and questions people type into search engines and AI assistants around a topic, then ranks those terms by how many people search them, how hard they are to rank for, and what the searcher actually wants when they type them. The output isn't just a list of words. It's a prioritized map of intent a site can build content against.
Key highlights
- 91.8% of all searches are long-tail keywords, specific, lower-volume phrases rather than short head terms.
- 58.5% of searches now end in zero clicks, which puts intent-matching and AI visibility on equal footing with traditional ranking as a goal.
- B2B companies using strategic keyword research report 702 to 1,389% ROI from SEO, per Whitehat's 2026 B2B keyword research guide.
- AI prompts run 25 or more words on average, per Practical Ecommerce's research, creating a long tail roughly four times larger than the one conventional keyword tools were ever built to capture.
- Roughly a third of AI prompts represent entirely new information-seeking behavior that keyword tools will never see, since AI platforms don't publish query volume, demographic, or seasonal trend data the way search engines do.
Four kinds of intent behind a search
- Informational. The searcher wants to learn something, how to set up Google Ads, best served by a guide or explainer.
- Navigational. The searcher wants to reach a specific site or page, Google Ads login.
- Commercial. The searcher is comparing options before a purchase, best SEO tools.
- Transactional. The searcher intends to buy or convert right away, buy Ahrefs subscription.
Running the process
- Brainstorm seed keywords from the site's actual products, services, or subject matter.
- Expand each seed with a keyword tool to surface related terms and questions.
- Pull the metrics for each term: search volume, difficulty, cost per click, and intent.
- Prioritize by realistic opportunity and actual business value, not raw search volume alone.
- Cluster the prioritized keywords by topic and map each cluster to a specific page or content plan.
Keyword research once AI enters the picture
Keyword research now has to serve two audiences: traditional search engines ranking individual pages for specific queries, and AI systems synthesizing answers from much longer, more conversational prompts. Traditional keyword tools still supply the underlying volume and competition data, but they can't see the roughly third of AI prompts that represent genuinely new search behavior, since AI platforms don't publish prompt-level statistics the way search engines publish keyword data. The practical fix is treating keyword research as a starting point, then layering in prompt research and question-based content planning for the queries a keyword tool structurally can't see.
Frequently asked questions
What is keyword research?
Keyword research is the process of finding the words and questions people actually search around a topic, then choosing which to target based on search volume, ranking difficulty, and searcher intent.
What are the four types of search intent?
Informational, where the searcher wants to learn something; navigational, where they want to reach a specific site; commercial, where they're comparing options; and transactional, where they intend to buy or convert right away.
Is long-tail search really the majority of search volume?
Yes. 91.8% of all searches are long-tail keywords, the specific, lower-volume phrases rather than short, competitive head terms, which makes long-tail targeting a major share of realistic opportunity for most sites.
Do keyword tools still work now that AI search has grown?
They still supply the underlying volume and competition data, but they need to be paired with prompt research, since AI prompts average 25 or more words and roughly a third of them represent search behavior a keyword tool structurally cannot see.
What kind of ROI does good keyword research actually produce?
B2B companies using strategic keyword research report 702 to 1,389% ROI from SEO, reflecting how much targeting the right terms, rather than the highest-volume ones, drives real business value.
Why does zero-click search change how keyword research should be prioritized?
58.5% of searches now ending without a click raises earning an AI citation or featured placement to the same priority as ranking a blue link, which shifts prioritization toward terms where a full answer, not just a click, is achievable.