Keyword Clustering by Search Intent
Let ChatGPT group your keyword list by search intent and map each cluster to the right page.
Why we recommend itBrings order to keyword chaos, by real search intent.
Last updated: July 2026 · Collective Brain
Good for
- Turn a raw keyword export from Ahrefs or Search Console into clean topic clusters
- Decide per cluster whether to build a new page or expand an existing one
- Avoid cannibalization by mapping each search intent to a single page
The prompt
You are an experienced SEO strategist focused on search intent and site architecture.
Context:
- Website: [your domain]
- Topic or product: [your product]
- Raw keyword list (one keyword per line, optional search volume): [your keyword list]
Task:
1. Group the keywords into topic clusters. Combine terms that share the same search intent and the same expected answer, not just similar wording.
2. For each cluster, identify the dominant search intent: informational, commercial, transactional, or navigational.
3. Pick one primary keyword per cluster (the one with the clearest intent and best volume-to-effort ratio) and list the rest as secondary keywords.
4. Recommend a page type per cluster (guide, category, product page, landing page, FAQ) and map it to a concrete URL: new page or expand an existing one.
5. Flag any clusters that overlap in intent and risk cannibalization.
Output format: a Markdown table with the columns Cluster, Search intent, Primary keyword, Secondary keywords, Recommended page type, Page mapping. Below it, three short sentences on the biggest opportunities. Ask a follow-up question if the keyword list is too unspecific for a clean mapping.Replace the bracketed placeholders with your own details.
Frequently asked
How many keywords can I paste at once?
Stick to roughly 100 to 200 keywords per run. ChatGPT clusters cleanly at that size without trimming the list. For very large exports, split them across several runs and merge the clusters at the end.
Does ChatGPT detect search intent reliably?
ChatGPT infers intent from the wording and usually gets it right, but it does not see the live SERPs. Spot-check borderline clusters in Google Search before you commit the mapping to your site structure.
Related
Rather have it done?
Prompts are a start.
Results are our job.
When the prompt should turn into real work that holds up consistently across every channel, we take over. Start free, finish professionally.