Okay, so there’s doing QA scoring on ✨ vibes ✨ and there’s scoring based on Science™️. I know I’m a nerd for getting excited about this: there’s actual science to explain why the mechanics of conversations fail. These mechanics don’t always show up in average scoring rubrics (or in a way that conflates things), so I wrote an article on it. https://supportresearchlab.substack.com/p/field-notes-clarity-relevance-informativeness