Managing Conversation Analytics Metadata
You can set conversational analytics metadata to start an AI chat session with Kyvos Dialogs. The metadata that you set can be exported or imported as needed.
You can set AI settings for a specific semantic model by:
Setting column metadata
To set the metadata for conversational analytics, perform the following steps.
From the Toolbox, click Semantic Models.
Select a semantic model from the list.
Click the AI Settings tab and then select the Columns Metadata tab.
Use the search bar to find a column by name or tag, or apply filters such as Dimension, Measure, Visible, or Empty.
Click a column (for example, FolderID or FolderName) to expand its details.
Under Tags, add relevant tags by clicking + Add Tag to get better results in case of similar columns.
In the Values Synonym Sheet field, select the appropriate synonym sheet (for example, Country or State) to map alternate values. This will help LLM to generate results as per the provided synonyms.
Enter or update the Business Context to provide a clear description of the column’s purpose and usage.
Click Hide if you want to exclude the column from AI interactions.
Repeat the above steps for additional columns as required.
Use Collapse All or Hide All to manage the view efficiently.
After completing all updates, click Save to apply the changes.
Generating metadata using Kyvos Copilot
To generate metadata using Kyvos Copilot, perform the following steps.
Click the Ask Kyvos Copilot link to generate business context and tags for semantic model's column.
Use the search box or apply filters such as Dimension, Measure, or Empty to locate the required columns.
In the Filter Columns list, select the required option to filter the required columns.
Select one or more columns by selecting the checkbox next to each column name.
Expand the column and provide the following details.
Under Tags, add relevant tags by clicking + Add Tag to get better results in case of similar columns.
Add Business Context for the selected columns.
If information already exists for tags and business context, review them for the selected columns.
In the Additional Input Prompt panel, enter any specific instructions or context to guide metadata generation (for example, business terminology or usage guidelines).
Additional options to perform:
Expand all: Click this to view detailed metadata for all selected columns.
Select all: Click this to choose all available columns.
Clear all Metadata: To remove metadata for all selected columns.
Click Generate and Review to allow Copilot to generate metadata for the selected columns.
Review the generated metadata and make any necessary edits.
Click Accept and then click Save to apply the updated metadata or Cancel to discard the changes.
Exporting metadata
The metadata that you set can be exported in the CSV file.
To export the metadata for conversational analytics, perform the following steps.
Click the AI Settings tab > Columns Metadata, and then click Export Metadata at the bottom of the page.
The Export Conversational Analytics Metadata dialog is displayed.
Click Download. The CSV file will be downloaded. In the CSV file, tags will be displayed as comma-separated values, and the key elements field value will be shown as ‘Yes’ or ‘No’.
Importing metadata
The metadata that you set can be imported to the CSV file.
To import the metadata for conversational analytics, perform the following steps.
Click the AI Settings tab > Columns Metadata, and then click Import Metadata at the bottom of the page.
The Import Conversational Analytics Metadata dialog is displayed.
Select the required metadata file from your local machine.
Click Import. The success message is displayed.