Drupal AI Webinar | AI Search & RAG

Overview

In this session, we’ll explore how Drupal’s search experience can evolve from traditional
keyword-based search to modern AI-powered semantic search. We will set up and compare two
approaches: standard keyword search and a semantic search system using the AI Search
module paired with Milvus vector database.

This is a hands-on session featuring live demos, configuration walkthroughs, and observability
insights. By the end, you’ll understand how semantic search returns more meaningful results
and how to enable conversational discovery through a chatbot built on site content.

What You’ll Learn

Understanding Search Models

• How traditional keyword-based search works in Drupal and where it falls short.
• How semantic search differs and how embeddings capture meaning and context.

Setting Up AI Search with Milvus

• Installing and configuring Milvus locally.
• Connecting Drupal AI Search to the vector database.
• Indexing content into vector embeddings.

Optimizing Semantic Retrieval

• Chunking strategies and why chunk size matters.
• Understanding and tuning RAG Retrieval Thresholds.
• Differentiating main content vs. contextual content during indexing.

Enhancing User Interaction

• Improving UI with the AI Search Block module & Search API Views.
• Introducing chatbot-based conversational search for follow-up questions and deeper exploration.

Who Should Attend

• Drupal Developers and Site Builders exploring AI-enhanced search.
• Architects evaluating RAG and vector search patterns in enterprise platforms.
• Teams modernizing content discovery and knowledge retrieval experiences.

Key Takeaways

By the end of the session, you will:

• Understand how semantic search differs from keyword search.
• Have a working setup of AI Search with Milvus on Drupal.
• Know how to tune chunking and RAG thresholds for better retrieval.
• Be able to extend the experience with conversational chatbot interfaces.

Format

• Type: Live Interactive Session
• Duration: ~60 minutes
• Includes: Demos, configuration walkthroughs, and Q&A