Research keywords with AI
Build a keyword shortlist around what people mean and what your site can offer, using available search metrics and clearly identified gaps in the evidence.
To research keywords with AI, begin with a decision you need to make. A shop choosing between a buyer’s guide and a category page needs to know whether a query signals investigation or a purchase. The keyword-research skill gathers candidates and compares intent, difficulty, volume and business relevance where those signals are available.
Existing rankings can change the priority. A page already receiving impressions may deserve refinement before a completely new article is commissioned. Search Console evidence belongs to your site; market estimates from Semrush describe a different source and should remain recognisable as such.
What it does
- Expand a seed topic into relevant variations and questions.
- Find queries for which the current domain already appears.
- Group phrases that could be answered well on the same page.
- Separate informational, commercial and transactional intent.
- Prioritise opportunities using evidence and realistic business value.
Install and use
- Install the skill and authorise the official Semrush MCP with an eligible account for live metrics. A Search Console export is optional. Without metric access, provide available terms and leave missing figures unknown.
- Ask, “Research watercolour brush queries. Group them by intent and suggest which belong in a buying guide and which suit our shop categories.”
- Review each cluster against your existing pages and the selected search market before assigning content.
FAQ
Source
Made by Semrush Inc. Original skill: keyword-research, published on GitHub under the Apache-2.0 licence.
View on GitHub