Drupal AI after the DriesNote Rotterdam: what you can use today and what comes next

# Overview At DrupalCon Rotterdam on September 29, 2026, Dries Buytaert presented three pillars to expand who can use Drupal: multilingual, JavaScript/headless, and AI. This brief focuses on the AI pillar and what was shown on stage, with concrete notes for site owners, editors, and module developers.

# The working demo you can try today Demo, called "Northmoor University," is a complete university website built to show AI features running on realistic content. The demo is publicly available at drupal.org/ai/demo.

Five practical demo features highlighted:

• Find weak content and fix it: the system scans programs, courses, people, news, events and policies for accuracy, tone, and inclusive language, then flags weak items.
• Fix it without the expert: editors can address flagged items directly instead of waiting for a subject-matter specialist.

• Drafts, not surprises: connected AI assistants can draft updates (for example, changing all pages that mention a newly promoted professor). Drafts are created but nothing is published without human review.

For agencies and site owners the demo provides a quick way to evaluate features and how governance and guardrails behave in practice.

# Developer features shown AI: an AI-assisted workflow can generate a content template for articles, validate output against Canvas Workbench, ESLint, and Canvas CLI, and push validated template files to the site. This lets front-end developers get first drafts automatically while relying on Canvas tooling for correctness.

# Roadmap and next steps Rosetta Sprint: the immediate next step is a four-day sprint with key maintainers to produce machine-readable descriptions of Drupal's capabilities. The stated goal is to make Drupal's data and capabilities easier to expose to AI agents, orchestration tools, frontends, and other systems.

Community momentum: Dries noted the Drupal AI Initiative has over 100 contributors so far. The initiative includes demo work, tooling prototypes, and training programs referenced elsewhere in the Drupal AI Initiative materials.

# What this means for you

• Try the Northmoor University demo to see features in context before planning a project. The demo shows both features and governance in action.

# Quick links mentioned in the keynote

• Drupal AI Demo: drupal.org/ai/demo
• Canvas tooling referenced: Canvas Workbench, Canvas CLI, ESLint
• Prototype modules: Simple OAuth, MCP Server, Tool API

# Bottom line Drupal demonstrated a working AI-backed site that emphasizes human review, grounded answers, and auditability. The developer prototypes point toward reduced integration friction, and the Rosetta Sprint aims to standardize how Drupal exposes capabilities to agents and tools.