Models

The Models tab lets you configure which semantic layer models, dimensions, and measures your agent can access.

Why Configure Model Access?

Not every agent needs access to all your data. By configuring model access, you can:

• Focus the agent - Limit access to relevant models for the agent's purpose

• Improve accuracy - Fewer options means less confusion for the agent

• Enforce governance - Prevent agents from accessing sensitive data

Configuring Access

Enable/Disable Models

Each model in your semantic layer appears as a card. Toggle the model on or off to control whether the agent can use it.

Dimension & Measure Access

Within each enabled model, you can further restrict access to specific dimensions and measures:

1. Expand a model card

2. Toggle individual dimensions on/off

3. Toggle individual measures on/off

The agent will only see and use the enabled components when answering questions.

No Models Available?

If you see "No models available", you need to create models in your semantic layer first:

1. Go to Models in the left sidebar

2. Create models from your connected data sources

3. Return here to configure agent access

Best Practices

• Start with minimal access and expand as needed

• Disable models that aren't relevant to this agent's purpose

• For sensitive measures (e.g., salary data), create a separate agent with restricted access

• Review model access when you add new models to your semantic layer