Query Expansion (20–30 Sub-queries)

Knowledge Graphs

Query ExpansionCascading keyword expansion

Query Expansion broadens a seed query into 20–30 sub-queries using LLM and SERP data: builds complete topical coverage of the central entity.

Query Expansion is a technique that broadens an initial seed query into 20–30 sub-queries using a combination of LLM (generating predictions) and SERP data (People Also Ask, Related Searches). The goal is building complete topical coverage of the central entity — it's not about one keyword, but the ENTIRE network of related queries.

In the semantic audit pipeline, query expansion works in several steps: LLM generates initial sub-queries, crawler collects PAA and Related Searches from Google, system tags each phrase (CONFIRMED/PREDICTED/SERP-ONLY), and the result is a list of 20–30 phrases ready for clustering. The 20–30 quantity is the optimal minimum; less gives incomplete coverage, more generates noise and increases API costs.

For example, seed query 'personal injury lawyer' → sub-queries: 'personal injury lawyer fees', 'how to file a personal injury claim', 'personal injury settlement calculator', 'personal injury vs insurance claim', etc.

In practice, generate sub-queries from different sources (LLM + PAA + Related) and deduplicate. Coverage from one source is always incomplete.

Source: AI Semantic SEO Expert, Robert Niechciał (sensai.io)