Browse documentation

Topics & Prompts

A Topic is a question category your buyers care about. A Prompt is one specific way they'd actually ask it. You track visibility at the prompt level — topics just keep things organized.

How Tracking Is Organized

Each brand you track (a business or product line) contains Topics, and each Topic contains Prompts. Think of it like folders and files — topics group related questions together so you can see patterns.

Brand: Example Brand (running shoes)

└ Topic: "Best trail running shoes"

├ Prompt: "What are the best trail running shoes for rocky ground?"

├ Prompt: "Which trail running shoes last the longest?"

└ Prompt: "Best waterproof trail running shoes under $150"

└ Topic: "Comfortable work shoes"

├ Prompt: "What do podiatrists recommend for standing all day?"

└ Prompt: "Most comfortable office shoes that look professional"

An example brand with two topics. Visibility is measured per prompt, then rolled up to the topic.

Topics vs Prompts

A Topic represents a buying intent or question category — something your potential customers research before purchasing. Topics are short labels that describe what the buyer is trying to learn.

A Prompt is a question we ask the AI assistants on your behalf, the way a customer would. It's written in natural, conversational language — the way a real person asks a question, not the way a marketer thinks about keywords.

Writing Effective Prompts

Prompts that read like real questions give you the most useful picture. In our experience the best ones are specific, conversational, and include context about what the person actually cares about.

  • Good: "What do podiatrists recommend for flat feet running shoes?" — specific, has expert framing, sounds like a real question
  • Good: "Best running shoes for marathon training under $150" — includes budget and use case
  • Weaker: "best running shoes" — too generic to show you much about a buyer's situation
  • Weaker: "running shoes sustainable eco material wool" — a keyword list, not a question anyone asks

Think about what your ideal customer would literally type after opening an AI assistant. That's your prompt.

Where You Manage Them

Open Prompt Volumes under GEO in the left panel. The page lists your topics on one side and the prompts in the topic you pick on the other. In the topic list you can add, edit (name, layer and rationale) and delete topics. A topic with active prompts cannot be deleted until they are archived or removed. In the prompt list you can add a prompt, and archive, restore or delete one, or many at once. To add many at once, import an Excel or CSV file from the template the page offers.

How Many to Track

Setup suggests topics and prompts for you to review. As a starting point we suggest a handful of topics with a few prompts each, then adding more as you find new angles. Quality matters more than quantity. The number of prompts you can track depends on your plan (see Plans & Limits).

Each prompt is checked on the AI assistants your plan includes, so even a small set gives you a visibility picture on each one (see How We Collect Data).

Archive, Don't Delete

  • Archive a prompt that no longer matters: it stops being tracked and you keep what was collected before
  • Rename a topic freely: it's only an organizational label
  • Add new prompts when you spot gaps in your coverage or launch new products
  • Review your prompt set now and then: buyer language changes, and your tracking should too