Get started with Meta Agent
Meta Agent is an intelligent agent for data asset management. It automatically inventories database metadata, generates business knowledge, and supports natural language-based asset discovery and data analysis. This topic describes how to get started with the asset inventory and asset Q&A features of Meta Agent.
Prerequisites
-
An Alibaba Cloud account is created and identity verification is complete.
-
Meta Agent (Database Edition or Data Lakehouse Edition) is purchased. For more information, see Billing.
-
The database instance is registered in AI-Native Database Service.
Procedure
Step 1: Go to Meta Agent
-
Log on to the AI-Native Database Service console.
-
In the left-side navigation pane, in the AI Data Assets section, click Meta Agent.
Step 2: Start asset inventory
-
On the Meta Agent page, in the asset inventory card, click Start Inventory.
-
In the dialog box, select the inventory scope. You can select by instance, database, or table. Confirm your selection to start the inventory.
-
Meta Agent automatically scans and parses the metadata, and generates business descriptions, SQL comments, and usage notes for tables and columns. The inventory process may take a few minutes, depending on the size of the database.
Step 3: Use asset Q&A
After the asset inventory is complete, go back to the Meta Agent page and click Asset Q&A in the asset Q&A card. A Q&A panel opens on the right side of the page. Ask questions in natural language in the conversation panel. For example:
-
"What columns does the orders table have?"
-
"Which tables store user information?"
-
"Write a SQL query to count the active users this month."
Meta Agent uses the knowledge generated during the inventory to understand business semantics and return accurate results.
Step 4: View and manage data assets
In the left-side navigation pane, switch between Business Catalog and Standard Catalog to view data assets organized by business or technical dimensions. You can register more data sources to expand the scope of asset management.
What to do next
-
For the full feature set of Meta Agent, including Data Lakehouse Edition and knowledge base management, see Meta Agent.
-
Use the data asset knowledge from Meta Agent with other agents such as Analytics Agent to improve analysis accuracy.