Search results pages have changed shape. AI Overviews, chat-based answer engines, and rich results now sit between a searcher's query and the traditional blue links, which means classic keyword research — find a phrase, check its monthly search volume, write a page around it — isn't enough on its own anymore. The tools people used to lean on for this (Overture's suggestion tool, the original Wordtracker workflow built around meta-search data) are long gone, but the underlying discipline they represented — understanding the exact language your customers use — is more important than ever.
Search Intent Is the Real Starting Point
Before you look at volume, sort your keyword list by intent:
- Informational — the searcher wants to learn something ("how does compound interest work")
- Navigational — the searcher wants a specific site or brand ("chase online banking login")
- Commercial investigation — the searcher is comparing options ("best project management software for small teams")
- Transactional — the searcher is ready to act ("buy standing desk under $300")
A keyword with modest volume but clear transactional intent is often worth more than a high-volume informational term, especially now that AI Overviews frequently absorb pure informational queries before a user ever clicks through to a website.
Where to Actually Find Keyword Data Today
Modern keyword research pulls from several overlapping sources rather than one master database:
- Search engine ad platforms (like Google Ads' keyword planning tools) for volume and competition estimates
- Autocomplete and "people also ask" data, which reflects real, current phrasing
- Your own site search and support ticket logs, which often reveal the exact words customers use that no third-party tool will surface
- AI chat logs and community forums (Reddit threads, niche forums, review sites) where people describe problems in their own words, before they've learned your industry's jargon
The Long Tail Still Wins, For a New Reason
The long tail of search — thousands of low-volume, highly specific queries — was always valuable because it converts well and faces less competition. Today it matters for an additional reason: specific, well-answered long-tail questions are exactly the kind of content AI answer engines pull from when constructing summaries and citing sources. Broad, generic content increasingly gets summarized and bypassed; specific, well-structured answers to narrow questions are more likely to get cited and clicked.
Building a Working Keyword Matrix
A practical process that still holds up:
- Start with a handful of core "seed" terms for your business.
- Expand each seed into related terms using autocomplete, competitor site language, and customer-facing data (support logs, sales call notes).
- Group the results by search intent and topic cluster, not just by exact phrase — modern search engines match on meaning, not just exact wording.
- Prioritize based on a mix of intent, estimated difficulty, and how directly the term relates to something you can actually deliver.
Put the List to Work
Once you have topic clusters instead of a flat list of phrases, the next step is figuring out which specific long-tail variations within each cluster are worth targeting individually versus folding into a broader page.
Next step: Feed your seed terms into the Longtail Keywords Tool to expand your list into the specific, lower-competition phrases your audience is actually typing — then group the results into topic clusters before you start writing.