AI's Impact on Production Audio Prep: Alts, Metadata, and Delivery?

Hey everyone, I've been spending a lot of time lately on a short film project, recording primarily with my MixPre-6 II and an MKH 416. We shot on an ALEXA Mini, and it’s a dialogue-heavy piece. I've always followed pretty standard practices for prepping audio for post, really focusing on careful organization of 'Alts' for options and robust 'Metadata' for easy searching, especially since I often do my own sound design later on. Lately, though, with all these new AI tools for noise reduction, spectral repair, and even auto-labeling, I'm wondering how this shifts the game. Are ‘Alts’ becoming less critical if AI can clean up nearly anything? Or does the ease of AI cleanup just mean a higher standard for 'Delivery' and even more meticulous 'Metadata' for quicker AI integration? I'm curious what workflows you all are seeing out there.