A path to doing this might look like: - Cron job scraping and committing - add a post commit check that runs a linter/some checks/tests - assign out fixes to these issues - ask for those fixes to be done using git - pick an area to focus on, and enforce coverage in that area. You can still blindly deploy here, but at least you know when it's broken
As noble a goal of testing before deployment is, sometimes it's an enormous amount of effort to change development practices, workflows, team mindset and company mindset. You can only handle some of these at a time so choose a combination of the low hanging fruit (maybe there's a separate component) and the highest impact (every time this breaks the site goes down)
It's not perfect, it's a step in the right direction.