My First AI Agent: From XML Chunks to a Living Tableau Dashboard

2026-07-09 · Source: Data Engineering on Medium · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Data Science & Analytics · Depth: Intermediate, long

Summary

An AI agent was developed to automate Tableau dashboard creation, addressing the repetitive nature of BI analysis. The architecture involves dissecting Tableau's ".twb" XML structure into data source, worksheet, and dashboard shell components. Templates with dynamic placeholders were created, alongside `agent_rules.md` for instructions and `config.json` for project-specific values like field names and colors. The agent executes in two phases: data preparation, filtering raw data (e.g., 459 rows to 204, 9 categories to 4) and creating a fresh `.hyper` extract; and dashboard assembly, where it fills XML templates and packages the `.twb` into a `.twbx` file. After 10 iterations, the agent achieved a deterministic run in under one minute, generating a complete dashboard with zero manual steps post-execution, except for publishing.

Key takeaway

For BI Analysts or AI Engineers struggling with repetitive dashboard creation, this agent architecture demonstrates a viable path to automation. By understanding the underlying file structures like Tableau's XML, you can design deterministic agents that significantly reduce manual effort and ensure consistent outputs. Focus on clear rules and externalized configurations to build reliable systems, freeing up time for deeper analysis rather than decoration.

Key insights

Automating complex UI tasks like Tableau dashboard creation is feasible by decomposing the underlying file structure and using rule-based AI agents.

Principles

• Decompose complex UI files into fixed and dynamic parts.
• Design agents for deterministic, rule-based execution.
• Separate agent rules from project-specific configurations.

Method

The method involves analyzing Tableau's `.twb` XML, creating templates with placeholders, defining explicit agent rules, and using a `config.json` for dynamic values, then executing data preparation and dashboard assembly phases.

In practice

• Analyze XML/JSON structures of target applications.
• Create modular templates with dynamic placeholders.
• Use `config.json` for project-specific parameters.

Topics

• AI Agents
• Tableau Automation
• Business Intelligence
• XML Parsing
• Dashboard Generation
• Data Engineering

Best for: AI Engineer, Software Engineer, Director of AI/ML

Related on AIssential

• Build AI-powered dashboard automation agents with NLP on Amazon Bedrock AgentCore — Multi-agent AI systems can automate complex, multi-step operational workflows by delegating tasks to specialized agents.
• A Formal Framework for Declarative Agentic AI in Business Process Analysis — A formal framework using AGO methodology and a BPKB enables precise, automated, and sound Agentic AI for business process analysis.
• From Dashboards to Decisions: How We’re Building AI Agents at RaySuite AI — AI agents transform user interaction by providing intelligent, contextual answers and taking action over complex systems.
• Building retrieval harness for enterprise agents — Effective enterprise agents require a retrieval harness combining LLM capabilities with robust indexing and diverse data access tools.
• Product development in the agentic era — Claude Managed Agents enable scalable, cloud-hosted AI agents for rapid prototyping and workflow automation.
• OpenAI unveils Workspace Agents, a successor to custom GPTs for enterprises that can plug directly into Slack, Salesforce and more — OpenAI's Workspace Agents enable persistent, code-executing AI workers across enterprise apps, evolving beyond session-based GPTs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Data Engineering on Medium.