Documentation

Discover how Improvado’s AI Agent for Marketing Data Governance turns rule creation into a conversational experience—automating data discovery, logic building, and deployment without SQL.

‍

The new AI Agent for Marketing Data Governance drastically simplifies how teams create governance rules.

Now, all you need to do is describe the rule in natural language — the agent automatically explores your database, identifies the right tables and metrics, builds the logic, and shows you a preview before deployment.

Once approved, the rule is created instantly and becomes visible on the Governance Dashboard within 15 minutes.
This new conversational flow eliminates manual setup, technical details, and SQL writing — making rule creation accessible to everyone.

How to Use

Step 1. Ask the Agent to Create a Rule

The agent automatically finds the relevant tables and metrics, without requiring you to specify column names or schema details. Example: “Create a rule to alert me when CPC for Google Ads campaigns exceeds $20.”

Step 2. Agent Exploration

The AI Agent explores the database structure to locate the correct fields (e.g., cost, clicks, impressions). It identifies the formula needed (e.g., CPC = cost / clicks) and builds the rule logic automatically.

‍Step 3. Preview and Validation

Before creation, the agent displays a preview showing which records would currently be flagged.
Example: “2 campaigns active — CPC $16 and $9, both below the $20 threshold.” You can approve or adjust the logic directly in chat.

‍Step 4. Define Parameters

• Choose run frequency (daily, hourly, weekly).
• Decide whether to enable immediately.
• Add notification recipients (including non-Improvado users via email).

‍Step 5. Define Parameters

After confirmation, the rule is created and appears on your dashboard. You can view rule details, SQL logic (if needed), and modify the setup anytime through the AI Agent.

Conclusions

The updated AI Agent turns rule creation into a fully conversational experience. It intelligently discovers the right data, builds logic, generates previews, and automates deployment — all without requiring SQL or schema knowledge.