🤖AI Assistant

"AI assistant" is in private preview. Contact [e-mail] to request access.

The Euno AI Assistant provides an intuitive way to interact with your data model using natural language. It can help you explore your data assets, answer questions about your metadata, and perform complex queries without having to learn EQL syntax.

What you'll learn

• How to use the AI Assistant to explore your data model

• Understanding the assistant's capabilities and limitations

• How to provide context for more accurate responses

• Working with resources in the conversation context

Using the AI Assistant

The Euno AI Assistant is designed to help you navigate and explore your data model using natural language. Instead of writing complex EQL queries, you can simply ask questions about your data assets, and the assistant will help you find the information you need.

Getting Started

You can access the AI Assistant from the main navigation menu. When you open the assistant, you'll see a chat interface where you can start asking questions. The assistant is designed to be conversational, so you can ask follow-up questions and provide additional context as needed.

Assistant Capabilities

The AI Assistant can help you with various tasks related to your data model:

• Searching data resources: Ask about specific data assets, their properties, or relationships between resources

• Exploring metadata: Get information about tables, columns, dashboards, and other resources in your data ecosystem

• Generating EQL queries: The assistant can create and execute EQL queries based on your natural language descriptions

• Resource discovery: Find resources based on specific criteria, such as dependency relationships or usage patterns

Example Queries

Here are some examples of questions you can ask the AI Assistant:

• "Find all Looker dashboards in the marketing folder that depend on Snowflake tables in the sales schema"

• "Which dbt models have not been queried in the last 30 days?"

• "Show me all the dashboards that John Smith has created"

• "What resources depend on the customer_orders table?"

• "Find all tables with PII data based on their descriptions"

Working with Resources in Context

The assistant keeps track of resources mentioned in your conversation, making it easier to refer to them in follow-up questions. Resources in the conversation context are displayed in a side panel that you can toggle by clicking the "Context" button.

Adding Resources to Context

Resources are automatically added to the context when:

1. You click on a resource link within the assistant's responses

2. You explicitly mention a resource by name or ID

3. You use the "Add to Context" button in the resource sidepane

Viewing Resource Details

When the assistant mentions a resource, it will be displayed as a clickable link. Clicking on the link will open the resource sidepane with detailed information about that resource, including:

• Basic properties (name, type, URI, etc.)

• Relationships with other resources

• Usage statistics

• Custom properties

Working with EQL

The assistant can generate and execute EQL queries based on your natural language descriptions. When the assistant creates an EQL query, it will display it in a code block that you can:

1. View to understand how your question was translated to EQL

2. Click to execute the query in the Data Model screen

3. Copy to use elsewhere or modify

For example:

type="dbt_model" AND has upstream((type='dbt_source' AND database_schema='raw'), 1)

Tips for Effective Use

• Be specific: The more details you provide, the better the assistant can understand your needs

• Use resource context: Refer to resources already in context for more accurate responses

• Ask follow-up questions: The conversation is stateful, so you can ask related questions

• Review EQL queries: Examine the generated EQL to learn how your questions are translated into queries

• Clarify ambiguities: If the assistant misunderstands, clarify your question with more specific terms

Limitations

The AI Assistant has some limitations to be aware of:

• It can only access information that is indexed in Euno's data model

• Very complex queries might need to be broken down into simpler steps

• It works best with resources and properties that have descriptive names and metadata

• Some advanced EQL features might not be available through natural language queries