Somebody should build software that reviews your Git commits and/or pull requests and provides you with a set of manual tests to run, step-by-step user journeys to replicate, etc. as part of your quality assurance process.
While building Poli, my co-founder and I regularly conduct manual tests before and after we release new features but, as the codebase gets larger and the product more sophisticated, some parts of the application become easier to neglect or forget about; parts that may be impacted by what we've built.
We have a large number of automated tests across the application, but we still need to sit down and do the hard work of considering the various user journeys, behaviours and expectations that already exist and may change based on the merged code.
As more of the grunt and repetitive work associated with building is automated and the technical barriers to creating software are brought down, anything that could make evaluating the quality of the build easier seems, to me, a good thing.