A few further issues to report for the Beta:
Handling of duplicates. Duplicate handling seems to be worse in the Beta than in the official release. My experience is that a lot of 100% congruent duplicates are missed (websites, in particular, seem to be problematic) and that semi-populated duplicates are regularly missed (that is, all fields that are populated are the same, but 100% of the fields may not be populated causing the library to look different whilst referring to the same paper).
Importation of abstracts. A significant percentage of my papers that have a functional DOI (that refers to a page with the abstract) and/or links to my university library record for the paper (that contains the abstract) are not automatically importing the abstract.
Auto-population of website meta-data. I feel like the Beta is doing a much better job of this than the official release on average. However, I still feel like a huge amount of web reference metadata is not being imported. How do I know? Since about 2012 nearly every popular website has been embedding copious amounts of metadata for SEO purposes and social media sharing purposes ( [Social media optimization with Yoast SEO • Yoast](https://yoast.com/social-media-optimization-with-yoast-seo/ - Yoast gives an explanation of what I’m talking about - they happen to be one of the oldest and most popular Wordpress plugins for SEO).
Auto-population of Australian statute and case law. As an Australian, I feel like the Beta is not following the stated convention for referencing (nor populating the references in such a manner). Detail on how to reference Australian law can be found here Cases and legal authorities | Style Manual and Acts of parliament | Style Manual. I can also provide some guides from our top universities on the matter if it’s of any interest to your team.