Drupal AI & machine learning

Drupal provides a flexible foundation for embedding AI into content, search and operational workflows. By combining structured content, granular permissions and open integration options, AI capabilities can become a governed part of the platform rather than a separate, bolted-on service.

Metadata, Teasers and Image Alt Text

AI can help editorial teams complete short but important publishing tasks, such as summarising an article, creating a concise teaser or drafting a meta description within a specified character limit. Suggestions can be generated directly within the Drupal editing interface, allowing editors to review and refine them before publication.

AI vision models can also suggest alt text for uploaded images, including images added in batches through Drupal’s Media Library. Human review remains important to ensure that descriptions accurately communicate the image’s purpose and context. Used responsibly, this can reduce manual effort while supporting more accessible content.

AI-enhanced Search

AI can enhance Drupal search by combining conventional keyword matching with semantic search, which identifies relevant content based on meaning and context rather than exact wording. It can also generate concise, grounded summaries that help users understand results and find the information they need more quickly.

Retrieval-augmented generation, or RAG, allows responses to be grounded in approved content from Drupal and other organisational sources. This can support conversational search, answer generation with links to source material and federated discovery across multiple websites, document repositories and business systems.

Personalisation and Recommendations

AI can help Drupal deliver more relevant content, services and recommendations based on contextual signals, user preferences and previous interactions. This might include suggesting related articles, relevant guidance, products, events or the next useful step in a user journey.

We design personalisation around clear user and organisational outcomes, with appropriate consent, privacy controls and safeguards. Recommendations can be tested and measured to confirm that they improve discovery and engagement rather than simply adding complexity.

Conversational Assistants

Conversational assistants give visitors another way to find and interact with information. Instead of navigating a website or searching with specific keywords, users can ask questions in natural language and receive answers grounded in approved website content and other trusted sources.

Drupal content can be combined with information from PDFs, office documents, spreadsheets, APIs and external systems. Responses can include links or citations so users can verify the information and continue their journey through the underlying content. Guardrails, permissions, and escalation routes can also be applied where accuracy or sensitivity matters.

Internal Knowledge Assistants

Internal assistants can help authorised users access organisational knowledge across policies, procedures, technical documentation and other approved information sources. Rather than training a new model on this material, the assistant can use permission-aware retrieval to identify relevant information and provide grounded answers.

Zoocha uses this approach to make its Business Management System documentation easier for team members to access, reducing the time required to locate and review information held in Confluence. Access controls and source references help ensure users receive role-appropriate information and can verify it against the original documentation.

AI-assisted Content Creation

AI can support content teams with first drafts, summaries, rewrites, translations, tone adjustments and content adapted for different channels or audiences. These capabilities can be integrated into Drupal’s editorial interface, allowing teams to use AI within their established workflows.

For organisations with formal approval processes, AI should assist rather than replace editorial judgement. Prompts, brand guidance, validation rules, and human review can be built into the workflow so generated content remains accurate, accessible, compliant, and consistent with the organisation’s voice.

Intelligent Workflows and Automation

Drupal AI can automate repetitive tasks such as classification, taxonomy tagging, translation, moderation, enquiry triage and content enrichment. AI steps can be incorporated into controlled workflows with clear conditions, approval stages and audit trails.

This approach combines AI flexibility with predictable business rules. It is particularly useful when teams want to improve efficiency without letting an AI system make consequential decisions without appropriate oversight.

AI Agents and Governed Actions

AI agents can go beyond generating content or answering questions by using approved tools to complete controlled actions within Drupal. Depending on the use case and permissions granted, an agent might help create structured content, update fields, organise taxonomy, support site configuration or coordinate tasks across connected systems.

Every agent should operate within clearly defined boundaries. Role-based permissions, human approval, logging, and validation can ensure actions remain secure, reviewable, and aligned with organisational policies.