Understanding prompts and prompt groups
Last updated: March 30, 2026
Prompts are the building blocks that power your workflows. To use them effectively, it’s important to understand how prompt groups and prompts work together.
This guide explains both — simply and practically.

What Is a Prompt?
A prompt is a single instruction you give LLMs to generate a specific output.
Think of a prompt as:
A question
A command
A template with variables
A structured instruction
Example Prompts
What is the best music streaming service right now?
Best music streaming options
Tell me cheapest music streaming apps
Each prompt produces one specific type of output.
What Is a Prompt Group?
A prompt group is a collection of related prompts organized together for a shared purpose.
Instead of running prompts one by one, prompt groups allow you to:
Keep related prompts organized
Run structured workflows
Standardize outputs across projects
Scale content generation
Think of a prompt group as a folder of prompts that work toward the same goal.
Why Prompt Groups Matter
Prompt groups help you to stay organized, standardize output and improve efficiency.
When to Create a New Prompt
Create a new prompt when:
You need a different output format
The tone or structure changes
The audience changes
The use case is different
You want to test an improved instruction
Avoid cramming multiple goals into one prompt — clarity improves results.
N.B. Trendos currently supports prompts only in English language, since we do believe that it provides most accurate results in LLMs.
More on custom prompts:
Custom prompts allow you to simulate real user questions and track how generative AI platforms respond to them. By creating and monitoring prompts, you can understand how often your brand, product, or topic appears in AI-generated answers and identify opportunities to improve your Generative Engine Optimization (GEO).
What Are Custom Prompts?
A custom prompt is a question or instruction that you configure inside Trendos to test how AI models respond.
These prompts mimic real user queries and are used to evaluate:
Whether your brand or product is mentioned in AI responses
How competitors are positioned in answers
What sources or reasoning the AI uses
The sentiment of mentions
In Trendos, you can add custom prompts in "Brand Insights" by clicking on "Prompt group" and selecting "+Add custom prompt".
What counts as a custom prompt?
One custom prompt needs a selected country and LLM provider.
Example:
If you create a prompt "What is the best music streaming service?" and you want to track it in two countries and two different LLMs it would count as 4 separate custom prompts.
Best Practices for Writing Effective Prompts
Follow these guidelines to get the most useful insights.
Mimic Real User Behavior
Use natural language questions rather than keyword phrases.
Example: What is the best streaming service right now?
Instead of: Best streaming service
Custom prompts are a core part of Generative Engine Optimization. By creating realistic AI queries and monitoring their responses, you can:
Track how often AI systems mention your brand
Understand how competitors are positioned
Identify content gaps that affect AI recommendations
Measure changes in visibility over time
Using a diverse set of prompts that reflect real user intent will provide the most accurate view of your brand’s presence across AI platforms.
Best Principles:
Keep Prompts Focused - One goal per prompt.
Be Specific - Clear instructions = predictable outputs.
Test and Refine
Treat prompts like living assets:
Improve wording
Adjust constraints
Optimize formatting
Small changes can dramatically improve output quality.