AI can generate almost anything, but the real challenge is ensuring it generates the right output for the right context. In real-world Drupal projects, AI needs to understand organisational terminology, brand guidelines, editorial standards, accessibility requirements, business rules, legal constraints, content structures and the specific context in which it operates.
This session explores how Drupal’s Context Control Center (CCC) can move AI context from scattered prompts and integrations into a governed, reusable platform layer.
Through a realistic Drupal AI workflow, we will demonstrate how to:
- Create reusable organisational context
- Apply global, scoped and conditional context
- Manage editorial, accessibility and business policies
- Provide agent-specific context
- Control context and token limits
- Support human approval and governance
- Debug which context an AI agent received and why
- Keep AI context maintainable as organisations evolve
Rather than treating every unexpected AI response as a model or prompting problem, we will explore how context selection, scope, conflicting rules and missing knowledge can be analysed using familiar application-debugging principles.
The session will demonstrate a practical architecture:
AI Capability + Context + Governance + Evaluation = Reliable AI Experience
Attendees will learn how Drupal can evolve beyond being a CMS with AI features into a governed knowledge and orchestration layer for organisational AI, and leave with practical patterns for designing, governing and operating AI-powered Drupal experiences at scale.