Skip to content
Start free trial

Guides · 4 minute read

How keywords and prompts work together

People can express the same need in many different ways. Keywords help us estimate interest in a topic. Prompts help us check whether AI mentions your business when someone asks about it.

Start with what your customer needs

A developer choosing tools for a new product might search for “Vue component library”. They might ask an AI assistant “Which Vue component library should I use for a SaaS dashboard?” or “What are the best alternatives to Vuetify?” The wording changes, but both questions can be relevant to a business selling Vue components.

Example relationship

vue component library

Which Vue component library should I use for a SaaS dashboard?

One keyword can relate to several prompts. A prompt can relate to several keywords. The match depends on the customer need, audience and intent.

What each measurement tells you

Monthly searches
A keyword's estimated average monthly search activity for the displayed market. Use it as one signal of interest when choosing topics.
AI demand
An estimate of how frequently a keyword occurs in questions people might ask AI. Our provider uses modeled question data, including Google People Also Ask. It is not a count of ChatGPT, Gemini or Claude conversations. Read the data definition.
AI visibility
Whether your brand appeared in the answers collected for your tracked prompts. The dates and model labels show what we actually checked. A single answer is a sample, and future answers can change.

Why we show related keyword demand

A prompt can be phrased in countless ways, so attaching a precise audience number to that sentence would imply certainty we do not have. Instead, open its related keywords to see the existing demand estimates alongside the connection.

Read each estimate separately. Related keywords can overlap, so adding their volumes together does not give a prompt's audience. A keyword estimate also does not measure every question with a similar meaning. Unknown means we do not have a measurement; zero means the provider returned zero for that keyword and market.

Use the relationship to choose useful work

  1. Check business relevance. Does this topic connect to what you offer and who you help? A small, specific buying question may be more valuable than a broad popular topic.
  2. Look at the intent. Buying, comparison and learning questions need different answers. Similar words alone do not make a good match.
  3. Read the collected answers. If competitors appear and you do not, inspect the recommendations and cited sources before deciding what to improve.
  4. Answer the customer's question well. Use keywords to understand the topic and prompts to understand the question. Repeating a phrase is not a substitute for useful, relevant content.

Scout suggests a few relevant connections and saves them for quick access. Remove any match that does not make sense for your business. Some prompts will have no suitable keyword yet, and inactive prompts remain saved without being checked weekly.