A hands-on companion to AI-powered keyword research focused on execution rather than theory. Provides the practical LLM prompts for seed expansion, intent classification, and semantic clustering, and a lightweight run-order for doing keyword research inside an assistant. Points to the canonical Research-and-Strategy guide for the strategic framework and to the SEO Prompt Library for the reusable prompts.
Read the strategy first. The full framework — how AI reframes keyword research around intent, semantics, and trends, the tool landscape, and where human judgment decides — lives in AI-Powered Keyword Research. This page is the execution companion: the actual prompts you run to do the work.
This file exists so the “using AI for SEO” toolkit has the keyword-research prompts alongside the rest of the workflow. It deliberately doesn’t repeat the strategy — it shows how to drive an LLM (Claude, ChatGPT, Gemini, or an SEO suite’s built-in assistant) through the job.
Run order
Keyword research inside an assistant follows the same shape whichever model you use: expand → classify → cluster → prioritize → validate. Do it in that order, keep the outputs in a sheet, and treat the model’s work as a draft your strategist signs off.
The three prompts that do most of the work
1. Expand a seed.
You are an SEO keyword strategist. Seed topic: "<topic>".
Generate 40 keyword ideas covering long-tail, semantic/related-entity,
and question-based variants. Return a table: keyword | type | why it matters.
Do not invent search volumes.
2. Classify intent across a list.
Classify each keyword below by dominant search intent
(informational / commercial investigation / transactional / navigational)
and suggest the content format that best serves it.
Return: keyword | intent | recommended format.
<paste keyword list>
3. Cluster into topics.
Group these keywords into thematic clusters, each mapping to a single page.
For each cluster give: cluster name, member keywords, the one primary keyword,
and the page type. Flag any keywords that overlap clusters (cannibalization risk).
<paste keyword list>
Keep the prompts you refine in the shared SEO Prompt Library rather than rewriting them each time.
The two guardrails that matter here
- Never trust invented numbers. Models will happily fabricate search volumes and difficulty scores. Pull real metrics from a keyword tool (Ahrefs, Semrush, Google Keyword Planner); use the LLM for language, intent, and structure only.
- A human prioritizes. The model surfaces options; your strategist decides which clusters are worth the content investment, per the canonical guide.
Keep going
- AI-Powered Keyword Research — the strategy
- SEO Prompt Library — reusable prompts
- AI for SEO Workflows — where this fits end to end
More Guides
Run disciplined SEO A/B tests in seven steps — one metric, two variations, randomized segments, run to significance, track, analyze the winner, and iterate.
Build a topic cluster in seven steps — select and score a pillar, validate it, map subtopics, align to intent, architect internal links, publish, and measure.
Prepare your site for AI search in five steps — content architecture, entity consistency, E-E-A-T, structured data, and machine-readable structure.
Get your content cited by AI in seven steps — answer capsules, link-free extraction, original data, digital PR, community presence, consistent messaging, and tracking.
A seven-step walkthrough for setting up Google Search Console on a new site — property type, DNS verification, sitemap, GA4 link, users, URL checks, and a monitoring routine.