Data Access

Control exactly what data Formula Bot can access in your connected databases. Configure table and column restrictions, and set up row-level security filters for sensitive data.

Why Configure Data Access?

Setting data access restrictions helps you:

• Protect sensitive data - Exclude tables or columns containing PII, financial data, or confidential information
• Improve accuracy - Constrain the AI to relevant tables for faster, more accurate results
• Maintain compliance - Ensure the AI only accesses data appropriate for your use case
• Enable multi-tenant filtering - Use row-level security to filter data by organization or user

Accessing Data Access Settings

1. Go to My Data in the dashboard
2. Switch to the Connectors tab
3. Find your SQL connector and click the menu icon
4. Select Data Access

Access Modes

When configuring a connector, you have two options:

Allow All Data

The AI can query any table and column in your database. Best for:

• Quick setup when all data is safe to query
• Development or testing environments
• Databases without sensitive information

Configure Restrictions

Choose specific tables, columns, and row filters. Recommended for:

• Production databases with sensitive data
• Multi-tenant applications
• Compliance-sensitive environments

Table Selection

Select which tables the AI can access:

1. Click Configure access restrictions
2. The system fetches your database schema
3. Use checkboxes to select/deselect tables
4. Use Select All or Deselect All for bulk actions
5. Search for specific tables using the search box

Column Selection

Control access at the column level:

1. Click the arrow next to a table name to expand it
2. Select/deselect individual columns
3. The table checkbox reflects partial selection (some columns selected)

Use Cases for Column Restrictions

• Exclude email addresses or phone numbers from customer tables
• Hide salary columns from employee tables
• Remove API keys or tokens from configuration tables

Row-Level Security (RLS)

Add filters to restrict which rows the AI can access. This is essential for multi-tenant data or when certain records should be excluded.

Enabling Row-Level Security

1. Expand the Row-Level Security section
2. Choose one or both filter types:
   • Shared Identifier - Apply the same filter across multiple tables
   • Single Table Rules - Create custom filters per table

Shared Identifier

Use this when multiple tables share a common identifier column (like org_id, organization_id, or tenant_id).

Setting up a Shared Identifier:

1. Enable Shared Identifier
2. Click columns that represent the same identifier across tables
3. Set the comparison operator (equals, not equals, greater than, less than, contains, or IN)
4. Enter the filter value

Example: If your tables have org_id columns and you want to filter to organization 123:

• Select customers.org_id, orders.org_id, products.org_id
• Set operator to =
• Enter value 123

All queries will automatically include WHERE org_id = 123 on each selected table.

Single Table Rules

Create custom WHERE clause filters for individual tables.

Adding a Single Table Rule:

1. Enable Single Table Rules
2. Click Add Filter Rule
3. Select the table and column
4. Choose an operator:
   • = (equals)
   • != (not equals)
   • > (greater than)
   • < (less than)
   • LIKE (contains)
   • IN (list of values)
5. Enter the filter value

Examples:

Table	Column	Operator	Value	Result
orders	status	=	active	Only active orders
users	email	!=	%@test.com	Exclude test accounts
products	price	>	0	Only products with prices
regions	country	IN	US, CA, MX	North American data only

Combining Filter Types

You can enable both Shared Identifier and Single Table Rules. When both are active, filters are combined with AND logic.

Example:

• Shared Identifier: org_id = 123
• Single Table Rule: orders.status = 'active'

Result: WHERE org_id = 123 AND status = 'active'

Filter Preview

When configuring RLS, you'll see a preview of the generated SQL WHERE clauses. This helps verify your filters before saving.

Saving Settings

1. Review your table, column, and RLS selections
2. Click Save at the top of the configuration
3. Settings take effect immediately for new queries

Changing Settings

You can modify data access settings at any time:

1. Return to My Data > Connectors
2. Click the menu on your connector
3. Select Data Access
4. Your existing settings will be pre-loaded
5. Make changes and save

Best Practices

Start Restrictive

Begin with minimal access and add tables/columns as needed rather than removing sensitive data later.

Use Meaningful Filters

RLS filters should reflect real access control requirements, not just convenience.

Test Your Configuration

After saving, run a few test queries in AI Chat to verify the AI can access the data it needs.

Document Your Choices

Keep notes on why certain tables or columns are restricted for future reference.

Review Periodically

As your database schema evolves, review data access settings to ensure they remain appropriate.

Troubleshooting

Tables Not Appearing

• Verify the connector credentials have SELECT permissions
• Check that the database/schema is accessible
• Try refreshing the schema

Queries Returning No Data

• Check if RLS filters are too restrictive
• Verify the filter values are correct
• Ensure the filtered column exists in the queried tables

Permission Errors

• The connector may lack access to certain tables
• Contact your database administrator to grant appropriate permissions

Next Steps

• Set up a Knowledge Base to provide additional context
• Connect a database if you haven't already
• Learn about AI Chat to start querying your data