Skip to main content

​
Understanding AI prompts vs. SEO keywords

Traditional search engine optimization (SEO) focuses on keywords and phrases that trigger ranked lists of links. In Generative Engine Optimization (GEO), prompts are natural language instructions that require AI models to synthesize information and generate a conversational response. While SEO is about appearing in a list, KIME tracks how your brand is integrated into the narrative of an AI’s answer. Prompts allow you to measure intent, context, and response generation. Unlike static keywords, a prompt can test complex user scenarios, such as “Compare high-performance gym tops for hot weather,” which requires the AI to evaluate your brand’s specific attributes rather than just matching a search term.

​
What is an “effective prompt”

To get the most accurate data from KIME, your prompts should mirror how real users interact with LLMs. Every prompt in your workspace should contain three core elements to ensure high-quality tracking data:
  3. The Intent: The goal of the query, such as seeking a recommendation, a comparison, or a technical specification.

​
Strategic prompt categories

KIME does not use a fixed category template because tracking needs vary by industry. Instead, we recommend building a framework that reflects your business structure and your customers’ journey. If you are a large sports brand, your category logic might follow your product departments:
  1. Product Categories: “Men gym tops”, “Men running shoes”, or “Women yoga gear”.
  2. Comparison Queries: Prompts that ask the AI to weigh your brand against specific competitors.
  3. Problem-Solution: Queries focused on user pain points, such as “How to prevent blisters during long runs”.
You can create and manage these labels in the Settings menu under Workspace Categories.

​
Managing your prompt library

KIME processes your prompt library every 24 hours at 00:00 GMT. This frequency ensures that your visibility scores reflect the latest model updates and training data.

​
Using AI-generated suggestions

KIME simplifies the setup process by suggesting prompts based on your root domain and homepage content. You can find these in the Proposed tab within the Prompts menu.
  1. Approve: Moves the suggestion to your active tracking list where it will begin generating data during the next refresh cycle.
  2. Reject: Removes the suggestion if it is not relevant to your current strategy.

​
Manual prompt creation

For specific campaigns or niche product launches, you can add prompts manually.
1

Open the Add Prompt modal

Click the Add Prompt button in the top right corner of the screen.
2

Input the query

Enter the natural language question you want to track. Be specific to capture high-intent AI responses.
3

Assign location and category

Select the country for the query simulation and assign it to a category for filtered reporting.
4

Save and track

Once saved, the prompt will appear in your Active list and be included in the next 00:00 GMT scan.

​
Data retention and limits

All data generated by your prompts is stored for two years, allowing you to perform long-term trend analysis and seasonal comparisons. You can view this history using the date filters on your main dashboard to track how your AI visibility evolves over time.