Why use GitLab Auto DevOps instead of hand-writing pipelines?

Auto DevOps exists to remove the upfront cost of writing a working CI/CD pipeline from scratch. For a fairly standard app (a common language, a conventional build process, deployment to Kubernetes), Auto DevOps can go from an empty .gitlab-ci.yml to a built, tested, scanned, and deployed application with almost no manual pipeline authoring.

It's most valuable in three situations: early-stage projects that don't yet know their exact CI/CD needs, teams standardizing dozens of similar microservices where writing a custom pipeline per repo doesn't scale, and teams that want security scanning (SAST, dependency scanning) enabled by default rather than opted into manually per project.

The trade-off is flexibility. Once a project's build or deployment process diverges from what buildpacks and the Auto DevOps templates assume, teams typically end up overriding specific stages with custom jobs anyway, so Auto DevOps works best as a fast starting point rather than a permanent solution for pipelines with unusual requirements.