This custom visual has the following two main configuration sections:

Config

This section covers all the provided configurations for the chatbot, along with model deployment settings available for users with upgraded licenses.

Semantic Model: This section contains the settings for establishing a connection with your semantic model.

Chatbot: This section contains the settings for customizing the behavior and information display of the Chatbot.

**Suggestions:** This section contains the customization options for the suggestions displayed in the welcome screen.

Language Model: This section contains the settings and connection to a user-provided language model deployment.

Premium: This section contains an alternative connection path for users with a dedicated backend.

Semantic Model

• Semantic Model ID: The identifier of the semantic model the chatbot will connect to. For more information about how to obtain this ID, check the troubleshooting guide.

Chatbot

• Reasoning: Whether the chatbot uses reasoning before providing an answer or not, activated by default. Deactivating reasoning may impact the agent’s accuracy for complex tasks.
• Show response stats: Whether the agent’s messages display time and token usage statistics. Response stats do not impact the agent’s speed or accuracy.

Suggestions

• Suggestion: The selected suggestion. The following configuration settings will apply to this suggestion.
• Title: Title displayed in the suggestion card.
• Description: Description displayed in the suggestion card, under the title.
• Prompt: The prompt that will be automatically included in the input bar when a user clicks the suggestion card.

Language Model

These options are only available for users with a Pro or Premium license. They allow you to configure a different language model as the source of the chatbot.

• Bring your own model: Switches from the provided model to your configured model deployment on activation.
• Model provider: The service that grants access to the LLM. Currently supported providers include OpenAI and Azure OpenAI deployments.
• API key: The API key to access the model deployment. Keep in mind these are stored in the report and visible to anyone who can make edits.
• Model / deployment: Your preferred language model.
• Embeddings model: The embeddings model which will process any uploaded documents. Keep in mind changing it while documents are already uploaded will require re-uploading them.
• Reasoning effort: A configuration setting native to the model provider, which controls the level of internal calculation performed. Select ‘None’ if your chosen model does not implement this setting.

Premium

These options are only available for users with a Premium license. They provide an alternative connection method to a dedicated backend, not dependent on Microsoft’s native token authentication.

• Tenant ID: The Entra tenant ID the key was issued for. It can be left empty when users are signed in.
• Premium license key: The key provided to your organization. Keep in mind the configuration values are visible to anyone who can edit your report. If you need help regarding your license key, please contact the Creative Data team.

Format

This section includes all settings related to the display and formatting of the chatbot.

**General:** This section contains formatting options to customize the chat user interface.

Charts: This section contains formatting options to customize the charts generated by the agent.

Tables: This section contains formatting options to customize the table views generated by the agent.