AI in Drupal development — code review, tests and migration risks
🎧 Listen to every article back to back
AI helps most in a Drupal project when a developer uses it to map code, prepare a test plan, find migration risks and support code review.
Drupal projects often have many layers: content types, views, permissions, modules, custom code, external connections and older decisions. AI can help identify initial patterns faster, but it cannot decide which option is commercially sound or technically responsible. We therefore use AI mainly to support Drupal audits, migration planning and automated test preparation.
Where AI helps
Understanding and mapping code. When a project arrives without documentation, AI can help a developer understand what custom modules do, which external integrations are in use and where the risky areas are. It does not replace reading the code, but it can shorten the initial mapping phase.
Checklists and processes. Migration, upgrades or security incident response all require a structured approach. AI can quickly produce a first checklist that a person then refines and confirms.
First drafts for tests. Writing Playwright and PHPUnit tests is often repetitive work. Based on a short description, AI can generate a test skeleton that the developer refines and runs. The result must always pass human review.
Code review support. AI can notice repeated patterns — unused variables, insecure SQL queries or outdated API usage. It supplements, rather than replaces, a human code review.
What AI does not do
AI does not decide alone what should change in a public website. AI-generated code, tests or conclusions must pass developer review. In critical places, the result must be run, tested and manually checked when needed.
AI does not know the client's business logic or history. It can only work with what it is shown — which is why context matters. A poor prompt produces a poor result even with a good tool.
Good AI use does not mean less responsibility. It means better workflows, faster analysis and more checkpoints before changes reach the public website.
How AI was used when building WebPro's own website is described in How WebPro built its website. Our AI in Drupal service page explains how we apply AI-assisted approaches in Drupal work.
Need Drupal help?
If the article describes your situation, you do not have to read everything first. A real person will help you choose the next step.