AI Visibility

Semantic Variants

Discover the internal fanout queries AI engines use when answering user prompts, and create content to get cited.

When a user asks an AI engine like ChatGPT or Perplexity a question, the AI doesn't search with that exact query. It breaks it into multiple internal "fanout queries" — different angles on the same question. Semantic Variants shows you these queries so you can create content that gets cited.

How It Works

  1. 1Run prompts through AI engines via the Prompts page — the system captures the fanout queries
  2. 2Navigate to Semantic Variants (from the sidebar or from the Prompts page)
  3. 3Browse variants grouped by prompt or view all unique variants
  4. 4Create content or track high-value variants as new prompts

Variant Types & Citation Rates

  • First queries (34.4% citation rate) — The initial search queries AI generates. Highest priority for content creation.
  • Middle queries (17.6% citation rate) — Refinement queries that explore deeper angles. A "Blue Ocean" opportunity.
  • Last queries (21.2% citation rate) — Final verification queries. Good for comprehensive coverage.

Two Views

  • By Prompt — Accordion-style grouping. Each prompt shows its fanout queries with brand mention indicators, word transformations (added/dropped/preserved), and action buttons.
  • All Variants — A deduplicated list of all fanout queries across prompts. Search, sort by A–Z, frequency, or mention status.

Actions

  • Track — Add a variant as a tracked prompt. Choose an existing topic or create a new one.
  • Generate — Create content (blog post, FAQ, comparison, tutorial, case study, or product page) optimized for that specific variant.

Pro Tip

Focus on "First" variants where your brand is NOT mentioned — these have the highest citation rate and the biggest opportunity for improvement. Use the Generate button to create targeted content.
semantic variantsfanout queriesai searchcitationcontent generationbrand mentionsword transformations