Skip to main content

AI Concierge: Additional Prompts

Learn how to use additional prompts for the AI Concierge to specifically control its behavior.

V
Written by Valerie Larissa Feichtmair

Additional Prompts in the AI Concierge

Under Additional Prompts in the AI Concierge, you can specifically control the AI's behavior and adapt it for different use cases.

Three predefined categories are already available to help you get started:

  • Communication Style

  • Limitations

  • Escalations

You can also create your own categories at any time if you need additional structure.


General functionality

To create a new additional prompt, first open the desired category. Then click “Add additional prompt”, assign a name to the prompt, and define the desired prompt content.

An additional prompt is a short instruction for the AI Concierge. It allows you to define how the AI should respond or behave in specific situations.

Within an additional prompt, you can also:

  • use Member Properties such as #FirstName

  • include Agent Actions such as @Support request


Conditions for Additional Prompts

You can also define conditions for each additional prompt.

This determines when a prompt should be active. This allows you to control the AI Concierge very specifically, for example:

  • only for specific Courses

  • only for specific User groups

  • depending on progress or access

This way, you can flexibly determine which instructions should be taken into account in which situation.


Categories in detail

1. Communication Style

In the Communication Style category, you define how the AI Concierge communicates with your Members.

This category is particularly suitable for keeping the AI's overall writing style and general behavior consistent.

For example, you can define whether the AI should respond in a casual, friendly, direct, or particularly professional manner.

2. Limitations

In the Limitations category, you define what the AI Concierge should not do.

This allows you to define clear boundaries and guidelines for the AI.

Typical examples include:

  • not covering certain topics

  • not providing recommendations outside the platform context

  • not responding to sensitive content

This ensures that the AI Concierge operates only within the framework you define.

3. Escalations

In the Escalations category, you define when a request should be forwarded.

This is particularly helpful when the AI Concierge reaches a point where an automatic response is no longer sufficient.

Typical use cases include:

  • a Member is dissatisfied

  • a support request arises

  • a question cannot be answered

In this category, you can use Agent Actions specifically to, for example:

  • inform team members

  • trigger external processes via Zapier or Make

This ensures that important situations are not overlooked and can be addressed in time.


If you have any further questions, please do not hesitate to contact our support team via the HelpCenter in the bottom left. We usually respond within 5 minutes on weekdays between 09:00 and 17:00.

Did this answer your question?