Evidence first, then prioritisation
We combine automated analysis with manual review of important templates and tasks. Findings explain the impact, evidence and recommended action, then separate urgent issues from worthwhile improvements and lower-value noise.