Skip to main content

General Overview: AI Concierge

A basic explanation of what the AI Concierge is and how it supports your users.

V
Written by Valerie Larissa Feichtmair

What is the AI Concierge suitable for?

The AI Concierge is a global assistant that is available to users at any time within the platform.

It is particularly suitable for:

  • quickly answering questions about content

  • providing users with guidance within the platform

  • serving as a central point of contact for support

Unlike AI Agents, the AI Concierge is not tied to a specific process, but can be used flexibly in different contexts.

Set up the AI Concierge

To configure the AI Concierge, navigate to AI Features → AI Concierge in the sidebar.

You will automatically be redirected to the Configuration section.

The layout is similar to that of the AI Agents:

  • Left side: AI Concierge settings

  • Right side: Live preview

This allows you to test adjustments directly and review the Concierge's behavior in real time.

Sections in detail

Design

In the Design section, you define the visual appearance of the AI Concierge.

Here, you can:

  • define question suggestions that users can click directly

  • customize the appearance of the chatbot profile (e.g., glow or shadow)

These settings help you visually adapt the Concierge to your platform and make it easier for users to get started.

AI Persona

The AI Persona section defines how the Concierge thinks, responds, and accesses content.

It is divided into four subsections:

1. AI Model

Here, you select the AI Model used for responses.

2. Personality settings

Here, you define how the Concierge communicates.

This includes:

  • response length

  • Tone of Voice

  • form of address

  • emoji usage

You can also define a boundary message. This is displayed when the Concierge cannot answer a question.

3. Content settings

Here, you determine which content the AI Concierge can access.

There are three options:

  • Accessible Courses only:
    The Concierge only accesses content that the respective user has access to.

  • All Courses with special knowledge tolerance (recommended):
    The Concierge can access all content, but takes the user's Access Permissions into account.
    Using question tolerance, you can define how many questions about content that has not been enabled may be answered.

    This option is particularly useful because:

    • users receive answers to their questions

    • while also creating an incentive to access higher-value content

  • All Courses:
    The Concierge has unrestricted access to all content on the platform.

You can also:

  • explicitly exclude specific content

  • specifically enable content even if it would not normally be accessible

These options are available regardless of the selected mode.

4. Additional content

Here, you can enable additional sources for the Concierge:

  • enable Web Search

  • upload documents

  • include Member Properties

This enables the Concierge to respond even more flexibly to different questions and take additional information into account.

Additional prompts

In the Additional prompts section, you can define additional rules and instructions that complement the AI Concierge's behavior.

These are organized into categories (e.g., communication style, restrictions, or escalations) and help you control behavior in a targeted way.

How additional prompts work in detail and how to set them up is explained in a separate article.

Simulate users

With the “Simulate users” feature, you can test the AI Concierge under realistic conditions before publishing it.

You can select different users to see how the chatbot behaves in different scenarios.

The feature helps you:

  • review the chatbot's behavior for different users

  • test additional prompts, including their conditions

  • ensure that content and logic are displayed correctly

This allows you to track exactly how the AI Concierge responds to different conditions.

This is particularly helpful for:

  • checking the chatbot's behavior for different users

  • testing additional prompts and their conditions

  • ensuring that content and logic work as intended

This allows you to check whether the AI Concierge works correctly in different scenarios before publishing it.

By simulating users, you ensure that your AI Concierge:

  • works consistently

  • responds correctly to different user profiles

  • is reliably tested before publication

This helps you avoid errors during live operation and provide a better user experience.

Publication

Once the AI Concierge has been fully configured, it can be published.

You can define who the Concierge should be available to:

  • Team members only → ideal for testing

  • Members with access → publication for platform users

This allows you to check the Concierge internally first before enabling it for everyone.

If necessary, you can restrict or revoke access again at any time, for example if adjustments need to be made.

The exact details of publication are explained in a separate article.

Chats & statistics in the AI Concierge

After you publish your AI Concierge, two key sections are available to you: Chats and Statistics.

These help you understand usage, review conversations, and optimize the performance of your Concierge.

Chats

When you select the AI Concierge in the sidebar, you will automatically land on the “Chats” tab.

Here, you will find a complete overview of all conversations that Members have had with the Concierge.

The section is divided into two parts:

  • Left side: List of all chats

  • Right side: Detailed view of the selected chat

In the detailed view, you can see:

  • the user who had the chat

  • the date and time of the conversation

  • the number of AI Responses

  • credits used, including costs

Using the three-dot menu in the top right, you can also:

  • permanently delete the chat

  • open the user (e.g., for further review or actions)

Statistics

In the “Statistics” tab, you get an overview of how your AI Concierge is being used.

KPIs at a glance

In the upper section, you can see the most important metrics:

  • answered questions

  • started chats

  • unique users

  • credits used (including costs)

These values are particularly helpful for keeping track of usage and costs.

Activity history

The activity history shows how usage develops over a specific period.

You can switch between different metrics:

  • answered questions

  • started chats

  • unique users

  • credits used

Adjust period

At the top right, you can set the period individually:

  • predefined periods (e.g., “Last 7 days”)

  • or a custom period via the calendar (start and end date)

This allows you to analyze exactly how usage changes over time.

Users with the most activity

On the right side, you can see the most active Members.

Here, you have two options:

  • use the arrow icon to jump directly to this user's chats (filtered view)

  • use “View all chats” to return to the complete overview

This allows you to analyze particularly active users and better understand how your AI Concierge is used.

If you have any further questions, please do not hesitate to contact our support team in the Help Center at the bottom left. We generally respond within 5 minutes on weekdays from 09:00 to 17:00.

Did this answer your question?