Keyword Research When Half Your Traffic Is an LLM

How to do keyword research in 2026: the same volume and intent basics, plus a new filter for whether a click is even available anymore.

The short answer

How to do keyword research in 2026: pull search volume and difficulty as before, then add a click-availability check, because Ahrefs' latest update found AI Overviews now cut click-through rate on the number one organic result by up to 58%, up from an isolated 34.5% effect the same research measured in March 2025.

Keyword research still starts where it always did: pull a keyword, check its search volume, check its difficulty, check that the intent behind it matches what you're about to write. None of that changed. What changed is what happens after a searcher sees the results, and that step now belongs in the research, not just the write-up afterward.

The two numbers that still matter

Search volume and keyword difficulty (from Ahrefs, DataForSEO, Semrush, or similar) are still the starting filter. High volume plus low difficulty is still the sweet spot. That part of keyword research has not moved.

The number that got added

What has moved is how much of that volume converts into a click at all. Ahrefs' February 2026 update to its AI Overviews study found that pages which used to rank in the top organic positions now see a 58% lower click-through rate when an AI Overview appears above them, up from a 34.5% drop when Ahrefs ran the same study in April 2025. Out of every 100 clicks a top-ranking page used to get, roughly 42 are left once an AI Overview sits above it.

That decline is not even across query types. SparkToro's 2026 analysis of Similarweb browsing data found 68.01% of all Google searches ended without a click, up from 60.45% in 2024, but split that by intent: informational queries were 74% zero-click, transactional queries only 31%. A query that triggers an AI Overview and also carries buying intent still sends clicks; a purely informational query that an AI Overview already answers in full often does not.

What this changes in practice

Add one step between pulling volume/difficulty and committing to write: search the exact keyword and look at what sits above the organic results. Three outcomes:

  1. No AI Overview, no answer box. Treat it like keyword research always worked. Volume and difficulty are still the whole story.
  2. AI Overview present, transactional intent. Still worth targeting. The click ceiling is lower than the raw volume suggests, but comparison and buying queries keep sending traffic even past an AI answer.
  3. AI Overview present, purely informational. The click may not come at all. The keyword is now worth pursuing mainly for citation, meaning the page should be written to be the source an AI Overview quotes and links, not just a page competing for a click that increasingly does not exist. That is a writing problem more than a keyword-selection problem; see answer engine optimization for how to structure a page so it's the passage that gets lifted.

None of this replaces the fundamentals covered in SEO for small business. It adds one filter on top of them, and a GEO audit checks a page against exactly that filter: whether it's structured to be citable when it can't be clicked.

Frequently asked

How do I do keyword research, step by step?

Pull a keyword's monthly search volume and keyword difficulty from a tool like Ahrefs, DataForSEO, or Semrush, check the search results to confirm the intent matches what you plan to write, then add one more step: check whether an AI Overview or a direct answer already sits above the results for that query. A high-volume, low-difficulty keyword is worth less than it used to be if nothing below the AI answer gets a click.

Does search volume still mean what it used to mean?

The number of monthly searches is the same, but a smaller share of them end in a click. SparkToro's 2026 analysis of Similarweb browsing data found 68.01% of Google searches ended without any click, up from 60.45% in its 2024 benchmark, a trend it attributes mainly to AI Overviews.

Should I stop targeting keywords that trigger an AI Overview?

Not automatically. SparkToro's research notes that branded searches, local businesses, and high-intent transactional or tactical queries still send clicks even as the overall zero-click rate rises. Check the SERP for the specific keyword rather than avoiding a whole topic because it sometimes shows an AI Overview.

What is the biggest keyword research mistake right now?

Ranking a keyword list purely by search volume and difficulty and skipping the SERP check. Two keywords with identical volume and difficulty scores can have completely different click-through ceilings depending on whether an AI Overview, a featured snippet, or a wall of ads sits above the organic results.

How does this change which keywords are worth writing for?

It shifts weight toward keywords where being cited inside the AI answer is realistic even without a click, and toward transactional and long-tail queries where a click is still likely. Both call for the same underlying discipline: a self-contained, quotable answer near the top of the page, which is the same standard behind answer engine optimization.

More in SEO for Small Business

See how AI describes your brand

A free GEO audit takes about 90 seconds and tells you what ChatGPT, Claude, Perplexity and Gemini actually say about you.

Run a free GEO audit →

Written by YaBaDo's AI agents and reviewed by a human before publishing. If you find something wrong, tell us and we will correct it.