What can you ask the Structural Agent to do?

You can ask the Structural Agent to perform a variety of analysis and configuration tasks. The following are just a few examples.

Task	Example prompts
Explore workspaces	Which workspace did I update most recently? Which workspace did I most recently generate data for? What's the most recently failed job in my workspaces? Give me a breakdown of the number of workspaces by data connector. How many workspaces don't yet have a completed data generation job? Are all of the name columns in my workspaces protected?
Navigate through the workspace	Display Database View. Show me Table View for the transactions table.
Ask questions about the workspace data	How many name columns were detected in the data? Are there any sensitive columns in the customers table? Are there any JSON columns?
Filter the workspace data columns	Show me all of the unprotected datetime columns. Show me all of the primary key and foreign key columns.
Set column sensitivity	Make sure that all datetime columns are marked as sensitive. Mark the currently displayed columns as not sensitive.
Ask for generator recommendations	Suggest a generator for the occupation column.
Configure column generators	Apply the Timestamp Shift generator to all of the datetime columns. Shift the dates by 1 day before or after the current value. Apply the Product List custom generator preset to the product column. Apply the Conditional generator to the seniority column. If the value of the column is less than 4, use the Random Integer generator to generate a replacement value between 0 and 4. If the value of the column is greater than or equal to 5, use the Random Integer generator to generate a value between 5 and 10. Apply the JSON Mask generator to the customer_summary column so that properties under name have the Name generator applied with the correct name type, the location properties have the Address generator applied with the correct location type, and the name property under work/properties has the Business Name generator applied. Apply the Regex Mask generator to the Product ID field to use the Random Integer generator to mask the numeric parts of the value.
Display sample data	Show me sample source and destination values for the last name column in the customers table.
Manage document-based data	Enable Document View for the customer_summary column. How many name fields are there across the database collections? Mark the occupation field as not sensitive. Apply the recommended generators to the fields under customer/name. Apply the Timestamp Shift generator to all fields that contain datetime values.
Manage subsetting	Is subsetting enabled? How many tables are in the subset? How large will the subset be based on the current configuration? What is the subset target table configuration? Should any tables be lookup tables? Enable subsetting and configure the transactions table as the target table. Use 5% of the records. Change the target table configuration to use records where the date is within the past 3 months. For the attendees table, filter optional records to include records where organization is test. Make vendor_type a lookup table. Enable subsetting and include tables that are not in the subset.
Create and configure post-job scripts	Create a post-job script to add demo users to the users table in the destination database. Enable the add demo users post-job script. Change the order of the post-job scripts to make the add demo users script run first. Delete the add demo users script.
Identify and resolve schema changes	Are there any schema changes that need to be resolved? Resolve the sensitive schema changes. Apply the Categorical generator to the new column. Dismiss the schema change notifications.
Get information about jobs and diagnose job failures	How many data generation jobs have run for this workspace? How many columns were de-identified during the most recent successful data generation? When was the most recent sensitivity scan for this workspace? How many sensitive columns did it identify? Why did the most recent data generation job fail? What's the solution to enable the job to complete successfully?
Compose a draft email message to Tonic.ai support. The Agent creates the draft message and then displays an Email support button that opens the draft message in your email client.	Can you create a draft email message to support to report the issue with access to workspaces?
Generate scripts and API calls Note that the Agent can only use the information available from this guide. It does not have access to the generated API reference.	Can you show me the API call for assigning the Email generator to the email address columns?

Note that Structural Agent requests are still subject to your global and workspace permissions. You cannot ask the Structural Agent to perform a task that you do not have permission to do.