AI Agents

mindzieDataDesigner includes three AI agents, each designed for a different level of experience and use case. All three agents are accessible from the toolbar at the top of the application.

Quick Reference

Agent Purpose Best For
DB Assistant Explore databases and write SQL queries Users who want to understand their data
ETL Assistant Build event logs with AI-driven planning and execution Experienced users who know what they want
AI Event Log Builder Step-by-step guided event log creation First-time users who want guidance

Which Agent Should I Use?

"I just want to look at my data" - Use the DB Assistant. It is a general-purpose SQL assistant with no process mining knowledge. It helps you explore schemas, write queries, and understand your database.

"I know what I want and I want the AI to build it" - Use the ETL Assistant. It interviews you about your goals, creates an implementation plan, and then autonomously builds the event log.

"I'm new to this and want to be walked through it" - Use the AI Event Log Builder. It guides you through 8 sequential steps, asking questions along the way and making sure nothing is missed.

How to Access

All three agents are available from the toolbar at the top of mindzieDataDesigner:

  1. Open a project in mindzieDataDesigner
  2. Look at the toolbar icons next to Browser, Build, and Backup
  3. Click DB Assistant, ETL Assistant, or AI Event Log Builder

Each agent opens in its own chat panel where you can interact with it using natural language.

Shared Features

All three agents share common infrastructure:

  • Security - An input guard screens all messages before they reach the agent, blocking prompt injection and dangerous SQL
  • Data Source Management - All agents can list, create, test, and switch database connections
  • Schema Exploration - All agents can browse tables, columns, relationships, and preview data
  • Query Execution - All agents can run SQL queries against your connected database
  • Playground - All agents can save and manage queries in the Playground folder

The ETL Assistant and AI Event Log Builder additionally share:

  • ERP Knowledge Base - A database of known ERP table patterns for automatic system detection
  • ETL Methodology - Shared process mining methodology for consistent event log construction
  • Validation Tools - Script validation, event log building, and quality checks

Agent Comparison

Feature DB Assistant ETL Assistant AI Event Log Builder
Schema exploration Yes Yes Yes
SQL query writing Yes Yes Yes
Data source management Yes Yes Yes
Process mining knowledge No Yes Yes
ERP system detection No Yes Yes
Event log building No Yes Yes
Script validation No Yes Yes
Implementation planning No Yes Yes
Guided step-by-step flow No No Yes (8 gates)
Autonomous execution No Yes Limited
Modify existing event logs No Yes No
Partial tasks Yes (SQL only) Yes No (must complete all steps)