The open-data and civic-engagement ART turns on an approved assistant but never trains Scrum Masters or coaches on allowed data, review expectations, or disclosure. Shadow tools and pasted datasets appear within two sprints. What integration step was missing?

Turning a tool on is not the same as integrating it into the ART. Without training on approved use, data limits, review, and disclosure, coaches invent local practices—including shadow tools and risky pastes of open-data sets that are not actually public. The RTE includes enablement in the rollout so the train shares one responsible pattern. Restricting the assistant to a single person, skipping walkthroughs, or publishing keys all worsen the problem.