Drupal in the AI Age: How Strategic Platforms Emerge from Customer Requirements

After more than 15 years as Drupal specialists, we see a recurring pattern: customer requests are often formulated in very technical terms, while the actual strategic requirements only become visible in the course of the discussions. For a long time, the answer to new trends was quickly found: “Yes, this can be implemented with Drupal.”

At the latest with the increasing maturity of AI technologies, however, this answer is no longer sufficient. The talk “How to sell Drupal in AI times” by Niels Aers (CPO at Drop Solid AI) at DrupalCon Nara last December captures this development very well. Customers today expect fewer experiments and instead clear statements about benefits, costs, and data responsibility. The following article summarizes the key ideas of this talk and places them in context. For the summary and structuring of the talk, I used Google NotebookLM as support.

The initial AI hype has noticeably subsided. Gartner describes this phase in the Hype Cycle as the “trough of disillusionment.” For many companies, AI is therefore no longer an end in itself, but a tool that must pay off and be controllable. This is precisely where the real strength of modern platform architectures begins, especially Drupal.

This article shows how typical customer concerns can be understood not as an obstacle, but as the foundation for a future-proof digital platform. Drupal is viewed less as a collection of features and more as a strategic foundation for sovereign, flexible, and economically viable AI-supported solutions.

1. From expectation to reality: what questions customers ask today

Anyone who talks about AI today is quickly confronted with very concrete questions. After a phase of exaggerated expectations, three topics have emerged that shape almost every conversation:

Practical value instead of experimentation

Many companies have tested initial AI projects, often without sustainable added value. Accordingly, the new expectation is clear: AI should solve real problems. Examples include a search that actually delivers relevant results, or functions that relieve editors in their daily work – such as automated summaries or suggestions for tagging articles. The focus is on pragmatic, measurable improvements.

Data sovereignty as a prerequisite

Especially in the European environment, the handling of data is a central issue. The transfer of sensitive content to external, non-transparent AI services is increasingly viewed critically. Customers expect clear answers as to where data is stored, who can access it, and how processing takes place. Data sovereignty is no longer an added benefit, but a basic requirement.

Costs and transparency

The desire to increase efficiency often stands in strong contrast to available budgets. Studies show that many marketing and IT departments are under considerable cost pressure. Proprietary platforms with difficult-to-calculate licensing and AI costs exacerbate this problem even further. What is needed are solutions with a transparent total cost of ownership.

These points define very precisely what a modern digital platform must deliver and at the same time provide a clear basis for argumentation in favor of open, controllable systems.

2. Structured content as the basis for AI

AI functions are only as good as the data basis on which they operate. This principle is not new, but gains much greater importance in the AI context. Drupal has long had a strength here that is often underestimated.

Why structure is decisive

In Drupal, content is modeled in clearly defined fields. In contrast to purely text-based editors, this creates a clean semantic structure. This was already a major advantage for search engines, for example when using schema.org metadata. The same principle applies to AI applications: structured content can be analyzed, classified, and further processed much more reliably.

Core strengths of Drupal

In the context of the Drupal AI initiative, four fundamental characteristics can be highlighted:

• Flexible data modeling for complex content
• Scalability, even with very large content inventories
• API-first architecture for easy integration of external services
• Large ecosystem of extensions and integrations

Drupal is therefore suitable not only as a CMS, but as a central content hub that connects various systems.

3. Open DXP: combining existing solutions sensibly

Drupal offers the possibility to develop almost any function yourself. In practice, however, this is rarely the most efficient approach. Many requirements, for example in marketing or personalization, are already very well covered by specialized open-source projects.

Connect instead of redevelop

Instead of rebuilding complex functions entirely in Drupal, a modular architecture is recommended. Drupal takes on the role of the central content system, while other tools are selectively connected.

Typical components of such an architecture are:

• Drupal for content and structure
• Mautic for marketing automation
• Unomi as a customer data platform

Advantages of this approach

• Content can be used across systems
• Development and maintenance costs remain manageable
• Teams focus on individual requirements instead of standard functions

The technical complexity of multiple systems remains, but can increasingly be abstracted.

4. AI as the connecting layer

Composable architectures bring flexibility, but often confront users with new challenges. Different interfaces and workflows can make everyday work more difficult. AI offers an interesting solution here.

Agents instead of interfaces

Instead of manually switching between systems, AI agents can act as an intermediary layer. Users formulate a task, for example creating a campaign, and the agent carries out the necessary steps in the background.

Humans remain part of the process

In the short to medium term, a hybrid approach will establish itself: AI supports preparation and automation, while humans review, refine, and approve decisions. This interaction is realistic, accepted, and already feasible today. AI does not replace humans, but enables them to work faster and more efficiently.

5. AI applications with direct added value

To create acceptance, AI functions must be concrete and understandable. Some practical examples show how this can be implemented:

• Internal AI search (RAG) to relieve service teams
• Automatic persona creation based on real usage data
• Editorial assistance functions, such as summaries, tone adjustments, or taxonomy suggestions
• Translations directly in the CMS workflow (for example with DeepL)

These functions can be introduced step by step and quickly deliver visible results.

6. Thinking about architectural decisions in the long term

When choosing a DXP strategy, companies typically face three options: monolithic all-in-one solutions, fully custom in-house developments, or an open, preconfigured system based on open source.

The so-called “pre-orchestrated open core” approach combines stability with flexibility. Customers start with a manageable system and expand it as needed. License dependencies are avoided, and data remains under their own control.

Conclusion

AI is changing not only technical systems, but also the role of agencies and consultants. What is needed are fewer feature lists and more well-founded architectural and strategic decisions.

Drupal provides a robust foundation for this: open, structured, extensible. In combination with specialized open-source tools and pragmatically applied AI, platforms emerge that are sustainable in the long term – technically and economically.

The focus therefore clearly shifts: away from short-term hype topics, toward sustainable digital strategies.