Drupal AI Initiative: Your next website visitor might not be human

For most of the web’s history, we have designed digital experiences around a simple assumption: a person will visit our website. That person might arrive through a search engine, follow a campaign link, scan a QR code, or maybe even type the URL into their browser.

AI is changing that... dramatically and rapidly!

People are now asking AI assistants to research products, compare services, explain policies, recommend suppliers and complete tasks on their behalf. Sometimes, they might not even consciously choose AI and are simply guided by seemingly familiar tools like Google 'AI Overviews'. Either way, instead of visiting ten websites, a customer may ask one assistant to gather the relevant information and present a recommendation. 

In the near future, that AI assistant could be doing more than reading a web page: checking product availability, requesting information, preparing an application, arranging an appointment or even completing a transaction.

Your next website visitor may not be a person at all, but an AI agent acting on their behalf, which raises a serious question:

Can AI systems understand our organisation, trust our information and interact with our services safely?

Getting to know your new audience

To be useful, AI assistants need to find the right information, understand its meaning and decide whether it is current and trustworthy.

A prospective student asking an assistant to compare courses across several universities, a buyer requesting a shortlist of products that meet detailed technical, ethical and budget requirements - both are now part of your website’s audience.

While human visitors use navigation, page layouts, graphic cues and calls to action, AI systems depend more heavily on structured information, descriptive metadata, clear relationships and reliable access to data.

Your web pages may look perfectly clear to a person but remain ambiguous to a machine. For example, a human might understand from the design that one contact address is intended for media enquiries, and another is for customer enquiries, but an AI assistant may not interpret it correctly unless it’s represented clearly in the underlying content structure.

The content management decisions you make today will shape how accurately they are represented by AI tomorrow.

Being visible is not the same as being understood

Many organisations are currently focused on whether their content appears in AI-generated answers. That is important, but visibility is only one part of the problem.

An AI system also needs to understand:

• What the organisation offers
• Which information is authoritative
• When the information was last reviewed
• Which products, services or locations it relates to
• Whether regional or language differences apply
• What actions can be taken
• Which information is public and which is restricted

Without this context, AI assistants may rely on outdated pages, confuse similar services or combine information that was never intended to be used together.

Preparing for AI visitors therefore requires more than content. It requires a well-structured and reliably governed source of truth.

Drupal gives content meaning

Drupal treats content as structured information rather than a collection of web pages. A university course, for example, could have defined fields for qualification, fees and application route, rather than burying them in a block of text. That structure is what makes the same content usable well beyond a single page.

For a human visitor, Drupal assembles that information into an attractive and accessible page. For an AI visitor, the same structure makes the information easier to identify, compare and reuse.

You don't need to maintain one version of content for people and another for machines because Drupal allows the same governed content to serve websites, applications, search services and AI agents.

Drupal can become the trusted source behind AI answers

AI systems are powerful, but they are only as dependable as the information and context available to them. The idea of autonomous agents can quickly become uncomfortable when governance is treated as an afterthought: what happens if an agent uses sensitive information, makes an unsuitable change, or you simply can't tell why an action occurred?

Drupal can provide a controlled source of organisational knowledge. Its content model, taxonomy and relationship system describe what information means, not simply where it appears on a page, helping an AI assistant distinguish a current policy from an archived one, or a general contact address from a specialist enquiry route.

The Drupal AI ecosystem is developing capabilities to support this level of governance, including guardrails for requests and responses, observability and activity logging, controlled access to organisational context, provider-independent integrations, and human review and approval workflows.

This is especially valuable for large or complex digital estates, where information is created by multiple departments across different languages and regions.Drupal's advanced AI implementation and integration does not negate all risk from AI usage, but it does give you a stronger foundation for identifying and managing it

Put simply, AI makes content governance essential to digital communication.

Hi, I'm a machine, please can I come in?

Making content understandable is the first step. The next is enabling controlled action

Giving an AI agent access to your digital platform creates an obvious concern: what will it be allowed to see and do?

Drupal has long supported detailed roles and permissions, allowing different users to view, edit, approve or publish specific types of content.

The same principle can be applied to AI visitors. A useful agent may need to inspect content, search records, or carry out an action, but it should never gain unrestricted access to your systems, or expose private content simply because that content exists in the same system. It should only be able to access the information and tools permitted for the person, service or task it represents.

 The Drupal AI Initiative organises this work through two connected areas:

• Inside AI, which brings AI assistance into Drupal for editors, marketers and site builders
• Outside AI, which enables external AI agents and tools to connect to and act on Drupal

This changes the role of the content management system from being a 'human experience engine' to being a governed platform through which people, applications and AI agents can understand and interact with your organisation.

Design for people, prepare for agents

Human visitors are not disappearing. People will continue to value clear information, strong design, accessible services and experiences that feel relevant and trustworthy. However, they will increasingly use AI to navigate and make sense of the vast amount of information available to them.

AI readiness can look like a technology challenge, but an AI system cannot reliably represent your brand if the underlying content is fragmented, duplicated or poorly structured.

The organisations that adapt successfully will not choose between human-centred design and machine-readable content. They will build digital platforms that support both by creating information people can understand, data machines can interpret and processes agents can interact with safely.

Your next website visitor might not be human - will your digital platform know exactly how to help them?

Try Drupal today!