One of the interesting key points was the sheer number of groups that need to ingest these test results.
There's the product management team as a whole that need certain subsets of data.
The engineering team needs to know when things break due to code issues (features not working as intended).
The design team needs to know when certain actions aren't able to be completed (say due to a portion of a map being broken).
However, the results of tests are largely informational, NOT part of the main release process (i.e. all tests must pass or they absolutely cannot release). They aid the build process, not block it.
Automation results are still a key part of the release process, just not in a way that blocks developer workflow. There are far too many people contributing to each branch to stop all development when a test breaks.