Decomposition, Not Deflection: A Theory of the Automation Boundary in Customer Support

Hey ho science! I deposited my fourth paper for review this week, and this one's perhaps my favorite. It's called "Decomposition, Not Deflection: A Theory of the Automation Boundary in Customer Support". AI driven deflection rates are great and all, but what does it tell us about how many customers were actually helped? And what are we using as a foundation to determine what gets automated? The actual paper itself isn't publicly available yet (DM me if you'd like a copy though), but I did a breakdown of the contents on the Customer Support Research Lab substack: supportresearchlab.substack.com/p/decomposition-not-deflection