AI that works within your governance
The main risks of AI in government content are accuracy, bias, and auditability. We reduce them by using AI as an assistant, not a publisher: it drafts, suggests, and searches, and people review before anything goes live.
Drupal's structured content and role-based permissions give AI a well-defined, governed body of content to work from.
What you get:
- Faster content work. AI-drafted summaries, teasers, and tags that editors review and approve.
- Search that understands what people mean. Semantic search matches the meaning of a question, not only its keywords.
- Answers with guardrails. Chatbots that draw on your own content, with your data kept under your control.
Our approach to AI
Examples of our AI work
The “AI Edition” can be installed in GovCMS PaaS to take advantage of Drupal AI functionality and the site, on top of a wide range of AI services.
A feature we sponsored that lets site builders create and update Views using plain language.
Our module adds AI-generated summaries, teasers, and taxonomy tagging to the editing experience.
Frequently asked questions
The Drupal AI Initiative, launched in June 2025, is building AI capabilities into the platform. The 2026 roadmap covers AI-assisted page generation, context management so AI follows brand voice and governance rules, background agents for content tasks, AI search, design system integration, and audit trails. We're a Gold AI Maker in the initiative.
We use our Augmentor module to help editors draft, improve, and tag content. That reduces manual effort and helps teams keep quality consistent at scale. We also use AI to improve search, content recommendations, and automated tagging.
Yes, with appropriate governance. The Australian Government encourages agencies to explore AI while emphasising responsible and transparent use. In Drupal, AI works within defined roles and permissions, and editors can review changes before publication. We design our integrations with government compliance in mind.
Yes. We've worked extensively on search accuracy and relevance for government sites. Beyond keyword search on Apache Solr, we've implemented autocomplete, content boosting by type and recency, and faceted search by topic, audience, and date. We've also researched context-aware search, which adjusts results to the user's situation.
The main risks are accuracy, bias, and auditability. Poorly governed AI can produce inaccurate content or an inconsistent tone, which damages public trust. We implement AI as a drafting and assistance tool, not an autonomous publisher, and all AI-generated content goes through human review before publication.
Case studies
Latest insights