Drupal developers spend countless hours copying error messages from Watchdog, searching through Google, Stack Overflow, and drupal.org issue queues, trying to understand whether a problem comes from core, contrib, or custom code. What if you could get an AI-powered analysis with a single click?
In this session, we'll explore the AI Watchdog Analyst module — a contrib module that connects Drupal's built-in logging system to any AI provider through the Drupal AI module. We will install the module on a live Drupal site, trigger real errors, and see the AI analysis modal in action. We'll walk through the module's architecture, learn how to customize the system prompt for project-specific contexts, and discuss caching strategies for API cost control.
Key topics covered: integration with the Drupal AI module's provider-agnostic plugin system (OpenAI, Azure, Ollama, Gemini), automatic contrib module detection with issue queue context enrichment, Markdown-rendered responses, multilingual support, and 24-hour result caching. Attendees will leave knowing how to install, configure, and extend the module — and with the knowledge to build similar AI-assisted tools for Drupal.