> ## Documentation Index
> Fetch the complete documentation index at: https://chatobserver.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Prompts

> How prompts are created, grouped, scheduled, and maintained in Chatobserver.

# Prompts

Prompts are the foundation of Chatobserver. They define what the platform asks, where it asks it, and how often the question is re-run.

## What a prompt contains

A prompt can include:

* prompt text
* an optional description
* one or more target platforms
* a region
* a schedule
* personas
* prompt groups

Those fields determine both execution and how the data can be segmented later.

## Prompt statuses

Prompts can move through states such as:

* active
* paused or inactive
* archived

Use active prompts for your live monitoring set. Archive prompts when they no longer belong in the working catalogue but should remain in historical records.

## Scheduling options

Chatobserver supports multiple cadence patterns, including:

* manual
* daily
* weekdays
* weekly
* custom days
* recurring minute intervals

Choose the lightest cadence that still captures meaningful change. Not every prompt needs the same frequency.

## Groups and personas

Groups and personas help structure the prompt library without duplicating prompts:

* groups organize prompts by initiative, business line, market, or reporting lens
* personas organize prompts around audience types or use cases

Use them consistently. The value of filtering and reporting depends on disciplined tagging.

## Bulk import

If you are moving beyond a small starter set, use bulk import rather than hand-creating every prompt. The platform validates rows before import so you can catch schedule, group, and persona issues early.

## Good prompt design

Prompts work best when they represent concrete user intent. In practice that means:

* writing the way a real user would ask
* separating materially different intents into separate prompts
* avoiding giant catch-all prompts that are hard to analyze later
* keeping naming conventions consistent across the workspace

## Operational guidance

A healthy prompt library is usually:

* narrow enough to review
* broad enough to represent the category
* grouped clearly enough to compare slices of the business

If your team cannot explain why a prompt exists, it probably should not be on a recurring schedule yet.